Aditya Mishra
Let’s talk

Forward-Deployed AI & Product Leader

AI ambition. Meet execution.

I build agents that act, products people use, and teams that know how to ship them.

Product judgment. Hands-on engineering. A clear path from the customer problem to adoption.

Build with me Explore my AI work
5 platforms shipped in 10 months. Built as CTO at Focus Realm.
RetrieveReasonAct

Aditya MishraEngineer & CTPO

Where can
I help?

Your knowledge becomes a working advantage.

Agents that act. Answers grounded in your data. Voice workflows that take the next step. I build the product, the evaluation and the recovery paths around them.

How I’d start

Pick the job worth automating. Connect the knowledge. Put a working pilot in users’ hands.

Built: multi-tenant RAG voice agents and multilingual AI pipelines.

Choose how we work together.

A builder on your team.

Practical AI, full-stack delivery and product ownership—from the customer problem to adoption.

Put AI to work

  1. Pick the job worth automating
  2. Connect the knowledge
  3. Put a working pilot in users’ hands
Discuss this plan

The little things matter.
So do the big bets.

My foundation is computer science, design thinking and human–computer interaction. Usefulness, joy and adoption belong in the same product.

Small things done right compound. Bold ideas deserve a real test. I adapt the route without losing the purpose.

Learn.
Build. Adapt.Find the right problem

01 / 06 · Understand

Start with people, not a feature list.

I use design thinking to understand the job, the friction and what a useful outcome looks like. The experience should make life easier and feel good to use.

  • Talk to users and the people accountable for the result.
  • Define usefulness, delight and adoption together.
  • Make the constraints and success criteria explicit.

Focus Realm: stakeholder adoption came first, with WhatsApp-like ease as the usability benchmark.

Explore my experience
Understand → Make a proof → Land it

With an unclear brief, I spend more time listening and testing assumptions. A small working slice helps the team discover what deserves to exist.

Do one thing a little better.
Then build on it.

A clearer message. One less click. A safer fallback. A better handover. Each improvement makes the next one more valuable.

Small improvements compound over timeTwenty cycles of one percent improvement reach 1.22 times the starting point.Starting point20 cycles later
1.22×the starting point
1.01²⁰

Illustrative: 1% compounded per cycle.

Design thinking. Hands-on engineering. Stakeholder adoption.My design roots ↗

Intelligence,
put to work.

Generative AI, LLMs, RAG and LangChain.
Here’s where capability becomes something you can use.

5LMS products
10months of delivery
120+client locations

Focus Realm · Education & hospitality

Voice Agent Labs2025

Your knowledge. On the line.

Built a multi-tenant AI voice platform: each business brings its knowledge, agent personality and call goals. The agent works across sales and support.

RAG, call evaluation, queue optimization and outage recovery—in one delivery.

  • RAG
  • Voice agents
  • Evaluations
  • Pub/Sub
Impact & architecture

Drove 3–4× outbound and 8–10× inbound telecalling throughput across production queues, backed by automated post-call evaluations and graceful fallback routing when external providers degraded.

See the work

RAG & voice

From input to outcome

6 seconds per step · Continuous loop

Step 1 of 4

Source knowledge

Select a step to explore the delivery.

Generative AI & LLMs

From model capability to a product people reach for.

RAG & LangChain

Connect your knowledge to the next useful action.

Voice, ASR & OCR

Let people speak, scan and work in their own language.

Product & adoption

Build the experience. Enable the team. Land it with users.

13+ years delivering excellence

Leadership you
can put to work.

International commerce. Enterprise learning. Voice AI. Startup launches.
Accountable for the product, the team and the outcome.

I build teams,
products & ecosystems.

I’m happy to take on challenges. Purpose over process. Results over excuses.

Since TexTimz (June 2013), I’ve combined MERN/FERN full-stack engineering with leadership across CXO, Fractional CTPO and delivery roles.

17 Leadership roles · 18 Engineering roles In 16 of those seats, I did both at the same time.

Selected roles

Aug 2025 — presentFyndSenior Solutions ArchitectEcommerce expansion, practical AI and customer delivery.Enterprise Ecommerce · Global & India Rollouts

Senior Solutions Architect

Aug 2025 — present

Leadership + engineering

Solutions architecture

Hands-on full-stack engineering across MERN and FERN.

The setting

International brands bringing ecommerce to India and global markets.

My remit

Connect customer goals with practical architecture, AI, automation and hands-on delivery.

What I bring forward

A customer-facing builder who can translate business needs into working systems.

  • MERN
  • FERN

Mumbai, Maharashtra, India

Impact & delivery
  • Help international brands launch and run ecommerce in India and worldwide.
  • Enable high-value clients to put AI and automation to practical use across ecommerce.
  • Use team leadership and hands-on skills to build solutions in days instead of months.
Oct 2024 — presentFocus RealmTechnology Advisor · Chief Technology Officer2 rolesLearning platforms built around stakeholder adoption.CTO chapter · 5 platforms shipped in 10 months

Technology Advisor

Aug 2025 — present

Leadership + engineering

Advisory

Hands-on full-stack engineering across MERN and FERN.

The setting

The advisory chapter following my CTO tenure at Focus Realm.

My remit

Continue as Technology Advisor, with the earlier learning-platform work providing the context.

What I bring forward

Continuity beyond a delivery engagement; the advisor and CTO chapters retain their own dates in the same company card.

  • MERN
  • FERN

New Delhi, Delhi, India

Impact & delivery
  • Continue supporting Focus Realm as a Technology Advisor following the CTO tenure that ended in August 2025.

Chief Technology Officer

Oct 2024 — Aug 2025

Leadership + engineering

Technology leadership

Hands-on full-stack engineering across MERN and FERN.

The setting

Learning products for schools, universities and hospitality groups, built around stakeholder adoption.

My remit

Lead product, engineering and QA across AI-assisted delivery, platform architecture and observability.

What I bring forward

Technical leadership that connects shipping the product with helping people use it.

  • MERN
  • FERN

New Delhi, Delhi, India

Impact & delivery
  • Lead technology for an essentials-first learning platform designed to feel simple, reliable, and consistent for students, teachers, schools, and—over time—parents.
  • Use stakeholder adoption as the product benchmark: if someone can use WhatsApp to upload files, they should be able to use Focus Realm without training.
  • Center the product and operating approach on customer success and a Stakeholder Adoption First philosophy.
  • Shipped five AI-assisted LMS applications in ten months: four enterprise-grade products for 4- and 5-star hotel groups and medical and chartered-accountancy universities, plus one product for schools.
  • Acquired enterprise clients serving 120+ national and international locations with localized content.
  • Led roughly 10–12 developers, designers, and quality testers across React, Node.js, Firebase, and edge functions.
  • Set up product observability with PostHog, Hotjar, and Microsoft Clarity.
Nov 2023 — presentJoys of Work ITES Pvt. Ltd.Technology Advisor3 roles · product familyModern tools, generative AI and more joyful work.Product family · Brew for hiring, Daksh for focus

Technology Advisor

Nov 2023 — present

Leadership + engineering

Advisory · product family

Hands-on full-stack engineering across MERN and FERN.

The setting

A family of products aimed at making hiring and everyday work more joyful.

My remit

Advise across the product family, including Brew for hiring and Daksh for focused work.

What I bring forward

Product judgment across related user problems; Brew and Daksh retain their product context within this company card.

  • MERN
  • FERN

Noida, Uttar Pradesh, India

Impact & delivery
  • Advise a company whose mission is to make work joyful through modern tools, software, and generative AI.
  • Support its product family, including Brew for hiring and Daksh for focused day-to-day workload management.

Brew Joys

Technology Advisor

Nov 2023 — present

Leadership + engineering

Advisory · hiring

Hands-on full-stack engineering across MERN and FERN.

The setting

The hiring product within the Joys of Work family.

My remit

Advise a hiring experience built around preferences, recommendations and transparency for both sides.

What I bring forward

Technology choices grounded in the experience of employers and people looking for work.

  • MERN
  • FERN

Noida, Uttar Pradesh, India

Impact & delivery
  • Advise a product that aims to make hiring—and getting hired—more joyful through preferences, recommendation engines, and greater transparency for both sides.

Daksh

Technology Advisor

Nov 2023 — present

Leadership + engineering

Advisory · productivity

Hands-on full-stack engineering across MERN and FERN.

The setting

The focused-work product within the same Joys of Work family.

My remit

Advise a lightweight approach to priorities and workload, combining Pomodoro, Eisenhower Matrix and AI ideas.

What I bring forward

Turn a complex prioritisation problem into an experience people can use day to day.

  • MERN
  • FERN

Noida, Uttar Pradesh, India

Impact & delivery
  • Advise a lightweight product that helps freshers and experienced professionals focus on important and potentially urgent work by combining ideas from the Pomodoro technique and Eisenhower Matrix with AI.
Jul 2018 — presentThe Disruption ChallengeChief Executive Officer · Fractional CTPO2 rolesStartup ideas through design, engineering, launch and handover.Fractional chapter · approximately 18 affiliated startups

Chief Executive Officer

Jul 2018 — present

Leadership + engineering

Founder

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams.

The setting

Entrepreneurial education for inexperienced and early-stage Indian founders.

My remit

Found and lead the platform, including its progressive web app for founders and startup teams.

What I bring forward

Founder-level ownership of the product and the people it serves.

  • MERN
  • FERN
  • Android
  • iOS

New Delhi Area, India

Impact & delivery
  • Founded and lead an education platform designed to help inexperienced and early-stage Indian founders build and launch technology startups quickly.
  • Built and manage a progressive web app that gives startup founders and team members round-the-clock access to entrepreneurial education.

Fractional CTPO

Oct 2018 — Mar 2022

Leadership + engineering

Fractional · startup delivery

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams.

The setting

Hands-on product work across affiliated startups during the 2018–22 chapter.

My remit

Carry ideas through design, engineering and launch, then help replacement teams take ownership.

What I bring forward

Build the product and leave behind a team able to continue it.

  • MERN
  • FERN
  • Android
  • iOS

Delhi, India

Impact & delivery
  • Worked as a full-stack developer, Fractional CTPO, technical advisor, and technical architect across products for approximately 18 affiliated startups, from zero to one and onward into scale.
  • Covered product ideation, UI/UX, product engineering, product and marketing strategy, and initial digital marketing.
  • Helped startups recruit replacement teams so the products and operating knowledge could continue beyond the initial engagement.
Jun 2025 — Aug 2025Voice Agent LabsFractional CTPO · FreelanceWhite-label voice AI for agency sales and support operations.3–4× outbound · 8–10× inbound telecalling throughput

Fractional CTPO · Freelance

Jun 2025 — Aug 2025

Leadership + engineering

Fractional · applied AI

Hands-on full-stack engineering across MERN and FERN.

The setting

Agencies running inbound and outbound voice workflows for businesses in the US and Europe.

My remit

Build a configurable, multi-tenant platform spanning RAG, agent behaviour, call evaluation, queues and recovery.

What I bring forward

AI delivery that accounts for customer boundaries, constrained phone access and service outages.

  • MERN
  • FERN

California, United States

Impact & delivery
  • Built a white-label, multi-tenant inbound and outbound AI tele-commerce platform for agencies and their client companies, with voice agents spanning local and global accents.
  • Supported outbound sales qualification, inbound sales, and customer support for companies selling goods and services in the United States and Europe.
  • Enabled agency customers to configure RAG knowledge bases, agent personalities, self-provided or platform-purchased phone numbers, call-success goals, evaluations, and management information reporting.
  • Designed queue optimization and graceful outage recovery across Next.js, Express, Firebase, edge functions and Pub/Sub. Drove 3–4× outbound and 8–10× inbound telecalling throughput.
  • Prioritized constrained phone-number access and collisions between inbound and outbound calls, with continuity measures for AI, telephony, voice-model, and client-website outages.
  • Secured annual recurring revenue commitments before launch.
Sep 2022 — Aug 2025Socialpi.aiChief Technology OfficerA platform portfolio for creators, agencies and D2C brands.18 products · 100k+ monthly active users

Chief Technology Officer

Sep 2022 — Aug 2025

Leadership + engineering

Platform leadership

Hands-on full-stack engineering across MERN and FERN.

The setting

Products for creators, agencies and D2C brands across Socialpi.ai and Affipi.com.

My remit

Build applications and edge services, grow a cross-functional team and make architecture choices around scale and cost.

What I bring forward

Hands-on engineering alongside the operating judgment needed to grow a product portfolio.

  • MERN
  • FERN

Noida, India

Impact & delivery
  • Built and scaled products for creators, agencies, and D2C brands across Socialpi.ai and Affipi.com.
  • Built 18 products, including 12 React micro-frontends and six Node edge services, serving more than 100,000 monthly active users while controlling cost through data structure and stack choices.
  • Managed and grew a cross-functional team of developers, designers, quality testers, and marketers, with services designed for horizontal scale.
  • Built 12 FERN-stack applications as an individual contributor in 16 months.
  • Coordinated data-driven marketing with agencies, creators and brands, acquiring and serving 50,000+ users and 45 D2C brands in 100 days.
Sep 2024 — May 2025Contint.AiFractional CTPO · FreelanceMultilingual speech and text workflows, built end to end.Solo-built · ASR, translation and OCR

Fractional CTPO · Freelance

Sep 2024 — May 2025

Leadership + engineering

Fractional · voice and language AI

Hands-on full-stack engineering across MERN and FERN.

The setting

An enterprise white-label platform for multilingual detection, transcription and translation.

My remit

Own the AI logic, interface, infrastructure, delivery and operations as a solo builder.

What I bring forward

End-to-end responsibility for turning language-model capability into a usable service.

  • MERN
  • FERN

Hyderabad, Telangana, India

Impact & delivery
  • Solo-built an enterprise white-label ASR platform for multilingual language detection, transcription, and translation for national and international clients.
  • Owned the AI logic, interface, infrastructure, code, deployment, scale, and operations using Python and Colab, React and Node, Whisper ASR, Tesseract and Syncfusion OCR, React micro-frontends, and Google Cloud edge functions.
Mar 2022 — Sep 2022KHILADI ADDAProject ManagerGamerPe: Android, iOS and web delivery with a complete team.GamerPe · 50k users in 30 days

Project Manager

Mar 2022 — Sep 2022

Leadership + engineering

Delivery leadership

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams for GamerPe.

The setting

The GamerPe mobile and web product, delivered with engineering, design, QA, administration and support.

My remit

Lead cross-functional delivery across Android, iOS, React applications and the Node/MongoDB backend.

What I bring forward

Coordinate the complete product journey, including the people who test and support it.

  • MERN
  • FERN
  • Android
  • iOS

India

Impact & delivery
  • Led developers, designers, testers, administration, and support through a 100-day mobile and web product delivery for GamerPe.
  • Delivered Android and iOS applications, two React applications, and a Node.js and MongoDB backend.
  • Reached 50,000 users in 30 days with zero customer-acquisition spend and roughly 70% of users returning daily.
Jul 2018 — presentFeels Good CreationsChief Innovation OfficerWeb and mobile platforms for entrepreneurial education.Innovation leadership · flagship Disruption Challenge initiative

Chief Innovation Officer

Jul 2018 — present

Leadership + engineering

Innovation leadership

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams.

The setting

The innovation and education work behind the flagship Disruption Challenge initiative.

My remit

Lead developers and designers building web and app experiences for founders and early employees.

What I bring forward

Connect an organisation’s innovation mission with tangible products and useful learning experiences.

  • MERN
  • FERN
  • Android
  • iOS

New Delhi, Delhi, India

Impact & delivery
  • Lead innovation and the flagship Disruption Challenge initiative with a team of developers and designers.
  • Build next-generation entrepreneurial education platforms using web and app technology, including a progressive web app with client-side data collection, visualization, and analytics for entrepreneurs and early employees.
  • The goal is to help young startups become more capable, successful, and investment-ready.
Sep 2017 — Jul 2018AccentureApplication Development AssociateA new lending application for an international bank.Salesforce + mainframe · team and delivery ownership

Application Development Associate

Sep 2017 — Jul 2018

Leadership + engineering

Banking · application delivery

Hands-on full-stack engineering across MERN and FERN.

The setting

A new lending application for an international bank, through finance work at Accenture.

My remit

Build on Salesforce and mainframe systems. Lead teams and own the application’s delivery outcomes.

What I bring forward

Hands-on engineering and team leadership on a banking product.

  • Salesforce
  • Mainframe
  • MERN
  • FERN

Gurgaon, Haryana, India

Impact & delivery
  • Worked on a new lending application for an international bank through finance work at Accenture.
  • The application was built on Salesforce and mainframe systems.
  • Led teams and owned delivery outcomes for the application while holding the Application Development Associate title.
Apr 2017 — Sep 2017WINNING TECHNOLOGIESAndroid Team Lead → Head of EngineeringMobile delivery into broader engineering leadership.Head of Engineering in 4 months

Android Team Lead → Head of Engineering

Apr 2017 — Sep 2017

Leadership + engineering

Mobile team leadership

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams.

  1. JoinedAndroid Team Lead
  2. Within 4 monthsHead of Engineering

The setting

An Android delivery chapter covering three mobile applications.

My remit

Join as Android Team Lead, lead development of three apps and progress to Head of Engineering within four months.

What I bring forward

A rapid step from mobile-team leadership into broader engineering responsibility.

  • Android
  • MERN
  • FERN
  • iOS

New Delhi Area, India

Impact & delivery
  • Joined as Android Team Lead and led the development of three Android mobile applications.
  • Promoted to Head of Engineering within four months.
Jun 2013 — May 2017TexTimzAndroid Development Intern → Team Lead → CTOHands-on mobile engineering into technology leadership.Team Lead in 3 months · CTO in 6

Android Development Intern → Team Lead → CTO

Jun 2013 — May 2017

Leadership + engineering

Mobile engineering

Hands-on full-stack engineering across MERN and FERN. Led Android and iOS teams.

  1. JoinedAndroid Intern
  2. Within 3 monthsTeam Lead
  3. Within 6 monthsCTO

The setting

My Android-development journey began at TexTimz in June 2013.

My remit

Start as an Android Development Intern, become Team Lead within three months and CTO within six months.

What I bring forward

My first progression from mobile engineering into team leadership and technology ownership.

  • Android
  • MERN
  • FERN
  • iOS

New Delhi, India

Impact & delivery
  • Started as an Android Development Intern.
  • Became Team Lead within three months and CTO within six months.
Feb 2017 — Apr 2017The Disruption Challenge '17 (TDC'17)Event OrganiserA startup pitching stage for student founders.Student startup event · Amity University

Event Organiser

Feb 2017 — Apr 2017

Leader

Student startup community

The setting

A pitching event at Amity University created exclusively for student startups.

My remit

Organise the event around the needs of student founders.

What I bring forward

An early connection between building products, entrepreneurship and bringing people together.

Amity University, Noida, Uttar Pradesh

Impact & delivery
  • Organised a startup pitching event created exclusively for student startups.
Jun 2016 — Aug 2016Deft Infosystems Pvt LtdMEAN Stack InternMEAN-stack development alongside graphic design.Engineering and design internship

MEAN Stack Intern

Jun 2016 — Aug 2016

Engineer

Full-stack internship

The setting

An internship combining web development with graphic design.

My remit

Build web applications with the MEAN stack and contribute graphic design.

What I bring forward

An early bridge between engineering and visual craft.

Impact & delivery
  • Worked as a MEAN-stack developer and graphic designer.
Jun 2015 — Jul 2015The Campus ConnectGraphics Design InternVisual stories for a student network and its publications.Templates, posters and infographics

Graphics Design Intern

Jun 2015 — Jul 2015

Engineer

Editorial design internship

The setting

A student network with weekly blog and journal publications.

My remit

Create templates, posters, infographics and visual content that help readers navigate editorial subjects.

What I bring forward

Make information understandable through visual communication, before the later product-engineering chapters.

New Delhi Area, India

Impact & delivery
  • Created templates, visual content, posters, and infographics for a student network and its weekly blog and journal publications.
  • The visual work supported editorial subjects including the Vyapam scandal and the 1975 Emergency.

What I build.
How I explore.

45 products, tools and studies.
Explore shipped products, open source and ongoing experiments.

Refine projects All focuses & technologies

Choose a focus Collection totals

45 projects and studies

Local build

Auto-Tinker

Can the agent actually improve?

Local evaluation loops for Claude Code, Codex, and Antigravity. Fixed scoring, holdouts, and keep-or-discard experiments.

  • AI research
  • Python
  • Ollama
  • Git
  • Evaluations
  • Local inference
  • Agent orchestration
  • Agent systems
  • Research
Contribution & architecture

Built the experiment loop around fixed scoring, isolated changes and holdout checks, so an agent’s proposed improvement can be accepted or discarded with evidence.

Local models and coding harnesses propose and evaluate changes. Improvement is still being tested.

Python · Ollama · evaluations · holdout gates

Experimental evaluation harness; improvement is still being tested.

Local build

Atlas

Know your entire AI stack.

An evidence-graded inventory of AI tools, models, skills, usage, and coverage. Includes the retired Stack Radar work.

  • Agent tooling
  • Node.js
  • React
  • AI observability
  • Developer tooling
  • Systems inventory
  • Agent systems
  • Research
Contribution & architecture

Connects tool discovery with usage and coverage records, helping distinguish what is installed from what has actually been exercised.

AI model, tool and harness inventory, with usage and coverage analysis.

Node.js · React · local scanning · usage analysis

Public

INTELLIAI

Useful AI intelligence, every day.

A source-linked AI intelligence publication with dated editions, release tracking, practical actions, and caveats.

  • AI-assisted publishing
  • JSON Feed
  • Editorial workflows
  • AI intelligence
  • Source verification
  • Knowledge synthesis
  • AI products
Contribution & architecture

Turns release research into dated, source-linked editions with practical actions and caveats, making fast-moving AI developments easier to assess.

AI-assisted research and editing, with source verification before publication.

Editorial workflows · structured feeds · source verification

Open project

Public

Tinker News

Know it. Learn it. Use it.

An AI newsroom with action-oriented stories, primary-source research, local-model generation, and feedback experiments.

  • AI-powered
  • Next.js
  • Firebase
  • Cloudflare D1
  • Ollama
  • LLM workflows
  • Local inference
  • Source verification
  • AI products
Contribution & architecture

Connects acquisition, story generation and reader feedback with a guarded publication workflow. Fine-tuning experiments stay separate from proven product outcomes.

Local-model generation and fallback routes support the newsroom; feedback and fine-tuning are experimental parts of the pipeline.

Next.js · Firebase · Cloudflare D1 · Ollama

Public endpoint verified. Fine-tuning is an experimental workflow, not a proven improvement claim.

Open project

Prototype

BrowserOCR

Read the image. Keep it local.

On-device PaddleOCR inference, fuzzy text correction, and offline tuning experiments; a separate worktree explores LiteRT.

  • AI-powered
  • TypeScript
  • PaddleOCR
  • ONNX Runtime Web
  • Hono
  • Supabase
  • Cloudflare Workers
  • Computer vision
  • OCR
  • Mobile web
  • On-device AI
  • Privacy
  • Offline support
  • AI products
  • Research
Contribution & architecture

Combines browser-side text detection and recognition with fuzzy correction and user-scoped learning. Designed around mobile-size OCR models, offline caching and keeping images on-device.

Computer vision: PaddleOCR runs through ONNX Runtime Web. Optional edge correction receives text, while LiteRT is a separate research direction.

ONNX Runtime Web · PaddleOCR · LiteRT research

Photos stay in the browser. Optional edge correction receives text.

Public

FactsPack

Knowledge, packed for agents.

An authored, hash-chained knowledge format with a browser converter, validation, and benchmark research.

  • Agent tooling
  • Node.js
  • Browser JavaScript
  • Knowledge engineering
  • Format design
  • Validation
  • Open source
  • Agent systems
  • Research
Contribution & architecture

Designed the format and its browser conversion and validation tools, with an evidence trail around compact agent-readable knowledge.

Agent-readable knowledge with deterministic hash chains and validation. Compression benchmarks remain exploratory.

Format design · Node.js · Validation · Research

View source ↗
Open project

Claude Design · Prototype

Rally Pro Player App

A premium match-day companion.

Claude Design explorations for match logging, tournament entry, live scores, and tactile mobile navigation.

  • AI-assisted design
  • Claude Design
  • Mobile UX
  • Product design
  • Design systems
  • Design
Contribution & architecture

Explores match-day flows and navigation through Liquid Glass, Modernist and Organic design-system directions for a mobile player experience.

Claude Design supported product exploration. Prototype match data is illustrative; AI-generated design is not runtime AI in the released app.

Claude Design · Liquid Glass · Modernist · Organic

Includes the three supporting design-system studies. Prototype match data is illustrative; this is separate from the released Rally Pro application.

Claude Design · Prototype

FactStack Product Design

Code intelligence, made legible.

Mobile-to-desktop design studies for code exploration, dependency graphs, change tracking, and evidence-led alerts.

  • AI-assisted design
  • Claude Design
  • Code intelligence
  • Responsive design
  • Information architecture
  • Design
Contribution & architecture

Makes code exploration, graph navigation and change evidence readable across phone and desktop layouts.

Claude Design supported responsive interface studies. Displayed analysis and savings are illustrative, not measured product results.

Claude Design · codebase UX · responsive systems

Design prototype; displayed analysis and savings are illustrative.

Claude Design · Prototype

Rally Pro Website Redesign

The easy way onto the court.

Court Glass landing-page explorations for players, hosts, and spectators, with a unified sport-led visual language.

  • AI-assisted design
  • Claude Design
  • Web design
  • Product storytelling
  • Design systems
  • Design
Contribution & architecture

Uses a unified sport-led visual language to guide players, hosts and spectators toward the right next step.

Claude Design supported the Court Glass landing-page exploration. This design is separate from production application behavior.

Claude Design · web experience · Court Glass

Design exploration; separate from current production behavior.

Claude Design · Prototype

Stacked Campaign Design

A bookshelf with a point of view.

Instagram promotion graphics and a dark-romance visual direction for Stacked's reading and shelf-discovery product.

  • AI-assisted design
  • Claude Design
  • Art direction
  • Brand design
  • Social campaigns
  • Design
Contribution & architecture

Carries the reading product’s identity into Instagram layouts and a dark-romance campaign direction.

Claude Design supported campaign graphics and art direction. The work is promotional design rather than a product AI integration.

Claude Design · campaign art direction · social layouts

Public

Rally Pro

A court. A match. A community.

Racket-sports discovery, matches, tournaments, and operations. Ongoing work adds release checks and a visual product inventory.

  • No AI feature documented
  • Astro
  • React
  • Supabase
  • Realtime
  • Mobile web
  • Product engineering
  • Sports technology
  • Products
Contribution & architecture

Brings player discovery, match participation and host operations into one sports experience, with separate player and administration surfaces.

Sports operations and realtime participation.

Astro · React · Supabase · Realtime

Public product; latest local fixes remain separate from deployment.

Open project

Public

Stacked

Your real bookshelf, connected.

Physical shelves, reading progress, series discovery, and browser-based book scanning.

  • AI-powered
  • React
  • Supabase
  • Cloudflare
  • ONNX
  • Computer vision
  • OCR
  • Mobile web
  • On-device AI
  • Privacy
  • Products
  • AI products
Contribution & architecture

Connects a physical bookshelf to a private reading record through OCR, barcode and ISBN capture, while making shelf sharing an explicit choice.

On-device computer vision supports book scanning. The product combines OCR with structured book lookup and reading workflows.

React · Supabase · Cloudflare · ONNX

Open project

Public

facts+

Make product quality inspectable.

UI, accessibility, and performance review with cited findings and machine-readable reports.

  • Agent tooling
  • TypeScript
  • Cloudflare Workers
  • SARIF
  • Accessibility
  • Performance
  • Quality engineering
  • Evaluations
  • Agent systems
  • Research
Contribution & architecture

Makes quality findings traceable through static, runtime and assisted review tiers, with HTML and SARIF outputs that teams can inspect.

Supports assisted product review and machine-readable evidence. Deterministic checks and assisted findings remain separate review tiers.

TypeScript · Workers · Browser runtime · SARIF

Open project

Public

Factstack

A map of your codebase.

Architecture, dependencies, change tracking, and MCP access for developers and coding agents.

  • Agent tooling
  • TypeScript
  • Remix
  • MCP
  • Vitest
  • Code intelligence
  • Developer tooling
  • Knowledge engineering
  • Agent systems
Contribution & architecture

Turns code structure into navigable maps, dependency graphs and change views, then exposes the same evidence to coding agents through MCP.

Codebase analysis exposed to coding agents through MCP.

TypeScript · Remix · MCP · Vitest

Open project

Public

Mise

Every shift, five-star.

Phone-first hotel operations for staff, managers, and standards authors. Timed work, service evidence, and handovers.

  • No AI feature documented
  • React
  • TypeScript
  • Vite
  • Mobile web
  • Hospitality technology
  • Workflow design
  • Role-based UX
  • Products
  • Design
Contribution & architecture

Translates service standards into timed shift work, evidence and handovers for staff, managers and standards authors on phone-first screens.

An operations prototype for timed tasks, service evidence and shift handovers.

React 19 · TypeScript · Vite · Responsive UX

Public prototype; V3 work and hosted acceptance are tracked separately.

Open project

Local build

KnowOrc

Give knowledge a working memory.

A local knowledge explorer with provenance, hybrid retrieval, and governed agent workflows.

  • Agent tooling
  • Remix
  • TypeScript
  • PGlite
  • PostgreSQL
  • MCP
  • Hybrid retrieval
  • Knowledge engineering
  • Agent orchestration
  • Local-first
  • Agent systems
  • Research
Contribution & architecture

Combines private knowledge exploration with retrieval and governed agent workflows, retaining provenance so people can inspect where context came from.

Hybrid retrieval and MCP supply context to agent workflows. Raw knowledge is private by default in the local workspace.

Remix 3 · TypeScript · PGlite/PostgreSQL · MCP

Prototype

Destiny Engine

From fragments to living stories.

A local story workspace for characters, worlds, narrative evidence, and author-owned canon.

  • Optional AI
  • Browser JavaScript
  • Node.js
  • Ren’Py
  • Creative tools
  • Knowledge engineering
  • Local-first
  • AI products
  • Research
Contribution & architecture

Organizes characters, worlds and source fragments while preserving the author’s canon separately from proposals and model feedback.

Model feedback is optional. Deterministic extraction, narrative evidence and author decisions remain distinct parts of the workspace.

Browser JavaScript · Node.js · deterministic extraction · Ren'Py

Prototype

Web Game Forge

Build the world. Prove it plays.

Reusable web-game systems, playable vignettes, asset and performance checks, and a source-tracked knowledge layer.

  • Agent tooling
  • React Three Fiber
  • Three.js
  • Rapier
  • SQLite FTS5
  • MCP
  • 3D web
  • Game engineering
  • Performance
  • Local-first
  • Products
  • Research
  • Design
Contribution & architecture

Pairs playable web vignettes with reusable game systems, deterministic asset checks and performance gates instead of stopping at a visual demo.

Source-tracked knowledge and MCP support game development.

React Three Fiber · Three.js · Rapier · SQLite FTS5

Prototype

AMPS Studios

A studio you can play through.

A story-led studio site with product, brand, growth, and team brief paths, plus an accessible Canvas game.

  • No AI feature documented
  • Remix
  • TypeScript
  • Canvas
  • Static prerender
  • Interactive storytelling
  • Accessibility
  • Game engineering
  • Design
  • Products
Contribution & architecture

Designed distinct product, brand, growth and team brief paths, with an accessible Canvas game woven into the studio’s story.

Interactive web storytelling and a Canvas game.

Remix 3 beta · TypeScript · Canvas · static prerender

Prototype

Glassy Spring

Physics you can feel.

A motion study of spring timing, squash and stretch, scroll animation, and glass surfaces with accessible fallbacks.

  • No AI feature documented
  • HTML
  • CSS
  • CSS linear()
  • Scroll timelines
  • Animation
  • Interaction design
  • Accessibility
  • Design
  • Research
Contribution & architecture

Explores how spring timing, surfaces and scroll choreography change the feel of an interface, with CSS fallbacks for accessibility.

Browser-native spring motion and scroll-driven interaction studies.

HTML · CSS linear() · scroll timelines

Prototype

MarkHub

Capture inspiration. Keep its context.

An early browser capture extension and dashboard for visual references; broader adapters remain exploratory.

  • No AI feature documented
  • React
  • MV3
  • Firebase
  • Framer Motion
  • Browser extensions
  • Visual research
  • Workflow design
  • Design
  • Products
Contribution & architecture

Connects a Chromium capture extension to a reference dashboard, preserving inspiration and context while broader adapters remain exploratory.

A browser-extension prototype for capturing and organizing visual references.

React · MV3 · Firebase · Framer Motion

Prototype

Blog CMS

A small CMS. Real boundaries.

A Remix blog with post editing, sessions, and server-side security patterns.

  • No AI feature documented
  • Remix
  • SQLite
  • TypeScript
  • Full-stack
  • Authentication
  • Application security
  • Products
  • Research
Contribution & architecture

Builds the complete editing flow around sessions, post CRUD and input handling, including CSRF protection, password hashing and sanitization.

Conventional content editing with authentication and application-security controls.

Remix · SQLite · TypeScript

Prototype

Symphony

Give coding agents a conductor.

Experimental implementations of OpenAI's Symphony specification: isolated workspaces, retries, and orchestration status.

  • Agent tooling
  • TypeScript
  • Express
  • GraphQL
  • React
  • Agent orchestration
  • Reliability
  • Developer tooling
  • Agent systems
  • Research
Contribution & architecture

Explores dispatching coding agents into isolated workspaces with retry behavior and visible orchestration status, following OpenAI’s specification.

Orchestrates upstream coding agents. The specification belongs to OpenAI; these are experimental implementations of it.

TypeScript · Express · GraphQL · React

Based on OpenAI’s Symphony specification.

Research

Datastore Dossier

Choose the database with evidence.

A source-linked comparison of Firestore, Supabase, ClickHouse, and PlanetScale, including cost and architecture tradeoffs.

  • No AI feature documented
  • Interactive HTML
  • Database architecture
  • Technical research
  • Cost analysis
  • Research
Contribution & architecture

Frames database choices through cited cost, performance and architecture tradeoffs, giving teams an inspectable decision method.

Database architecture research comparing cost and performance.

64-source research · interactive HTML

Concept

HospitalMan

Designing emergency-department flow.

A product and architecture proposal for emergency-department operations. Concept work, not a deployed clinical system.

  • No AI feature documented
  • Systems design
  • Healthcare workflows
  • Product strategy
  • Architecture
  • Research
  • Products
Contribution & architecture

Maps emergency-department workflows to product scope, delivery phases and a compliance-aware architecture proposal.

An emergency-department product concept and architecture proposal.

Product strategy · systems design · prototype

Research

Hospitality Learning Research

From standards to daily service.

Hospitality product discovery, competitor analysis, SOP authoring, and operating workflow studies.

  • No AI feature documented
  • Product architecture
  • Hospitality technology
  • Technical research
  • Workflow design
  • Research
  • Design
Contribution & architecture

Connects competitor evidence, learning-product architecture and SOP authoring to the daily work of hospitality teams.

Hospitality workflow research, product scoping and learning-system architecture.

Product architecture · research · prototypes

Prototype

Covenant

Fair work, on the record.

A contract-work prototype for shared plans, delivery commitments, acceptance, and evidence-backed disputes.

  • No AI feature documented
  • Remix
  • Cloudflare
  • Supabase
  • Drizzle
  • Workflow design
  • Full-stack
  • Delivery governance
  • Products
Contribution & architecture

Makes plans, commitments, acceptance windows and dispute evidence visible to both sides of a freelance delivery engagement.

Delivery records and decision evidence in a product prototype. Human mediation remains planned.

Remix · Cloudflare · Supabase · Drizzle

Prototype; named human mediation remains planned.

Concept

Facts Pro

One source of truth for teams.

A product concept connecting intent, decisions, tests, security, and provenance across AI-assisted delivery.

  • Agent tooling
  • Remix
  • TypeScript
  • Knowledge engineering
  • Delivery governance
  • Product strategy
  • Agent systems
  • Research
Contribution & architecture

Explores how shared intent, decisions, tests and provenance could become one inspectable record for teams using multiple agents.

A concept for governing AI-assisted delivery. Proposed workflows and architecture are not presented as shipped capabilities.

Remix · TypeScript · provenance design

Concept

DeFacts Portfolio Concept

A playful portfolio direction.

An early Claude Design concept exploring a Friends-inspired portfolio, liquid-glass surfaces, and spring animation.

  • AI-assisted design
  • Claude Design concept
  • Art direction
  • Portfolio UX
  • Animation
  • Design
Contribution & architecture

Tests a playful portfolio direction through typography, glass surfaces and spring-motion ideas; the interrupted concept has no saved rendered canvas.

An early Claude Design concept. It is not counted among the four exported product-design prototypes.

Art direction · portfolio UX

Concept only; the interrupted project has no saved rendered canvas.

Public

AIWHYS

From intelligence to personal context.

A Cloudflare editorial mirror of INTELLIAI. The local UserBrain explores reading history, goals, and personalization.

  • No AI feature documented
  • Cloudflare Workers
  • JSON Feed
  • Browser storage
  • Edge architecture
  • Personalization
  • Privacy
  • Progressive enhancement
  • AI products
  • Research
Contribution & architecture

Preserves source links in a public editorial mirror. A separate local UserBrain explores reading history, goals and private browser personalization.

Publishes AI intelligence, but the mirror and local recommendation logic make no model request. UserBrain is local and undeployed.

Cloudflare Workers · JSON Feed · browser-local storage

The editorial mirror is public. UserBrain is local and undeployed.

Open project

Local build

Facts Open

Understand a repo without uploading it.

A local codebase scanner, interactive architecture viewer, AI snapshot, and MCP server.

  • Agent tooling
  • Node.js
  • MCP
  • Code intelligence
  • Open source
  • Local-first
  • Privacy
  • Agent systems
Contribution & architecture

Lets developers scan a repository locally, browse its architecture and share an AI-ready snapshot without uploading their source.

A deterministic scanner and MCP server supply codebase context to agents. The open-source edition does not require model inference for scanning.

Node.js · local-first analysis · MCP

Open-source edition identified in the project README; public repository link is not independently verified.

Local build

Facts Ledger

One history across the facts family.

A local repository and session ledger joining Git state, project variants, and activity across three harnesses.

  • Agent tooling
  • PowerShell
  • Git
  • Provenance
  • Developer tooling
  • Cross-harness workflows
  • Agent systems
Contribution & architecture

Joins repository variants, Git state and session activity in an append-only local record across three coding harnesses.

Local provenance and activity tracking for AI-assisted development.

PowerShell · Git · append-only records

Local build

Agent Directory

Every harness. One work history.

A unified, local session index for Claude Code, Codex, and Antigravity, with incremental collection and recurring digests.

  • Agent tooling
  • Python
  • SQLite
  • Data indexing
  • Cross-harness workflows
  • Automation
  • Agent systems
Contribution & architecture

Collects new session records incrementally into one searchable local work history, with recurring digests across Claude Code, Codex and Antigravity.

Local indexing and deduplication of coding-agent sessions across harnesses.

Python · SQLite · incremental indexing

Local build

GBrain Broker

Many agents. One shared brain.

A local MCP broker that lets multiple harnesses share one PGlite-backed gbrain process without competing for its lock.

  • Agent tooling
  • MCP
  • Named pipes
  • PGlite
  • Systems integration
  • Knowledge engineering
  • Concurrency
  • Agent systems
Contribution & architecture

Coordinates access to one shared local gbrain process so different harnesses can use its knowledge without competing for the database lock.

MCP and named-pipe integration connects agents to shared knowledge. Upstream gbrain remains a separate project.

MCP · named pipes · PGlite integration

Integration work around gbrain; upstream gbrain is a separate project.

Local build

UI Inventory

See every screen and every state.

A browser capture foundation for screens, overlays, windows, themes, and responsive breakpoints, with portable reports.

  • Agent tooling
  • Node.js
  • CDP
  • Browser automation
  • Quality engineering
  • Responsive design
  • Accessibility
  • Agent systems
  • Research
  • Design
Contribution & architecture

Captures screens, overlays, windows, themes and breakpoint states in a portable report so interface reviews start from observable evidence.

Agent-led interface capture and inspection. Capture infrastructure is working; audit engines remain in development.

Node.js · CDP · portable agent skills

Capture foundation verified; this does not claim completion of all audit engines.

Research

Remix Research Labs

Read the source. Test the contract.

A framework wiki and exact-version labs for routes, forms, frames, import maps, and evolving runtime contracts.

  • No AI feature documented
  • Remix
  • TypeScript
  • HTTP
  • Technical research
  • Contract testing
  • Full-stack
  • Research
Contribution & architecture

Turns framework-source reading into exact-version examples and HTTP checks, making changing routes, forms and runtime contracts testable.

Version-pinned framework experiments and executable HTTP contract checks.

Remix · TypeScript · HTTP contract checks

Research

Longform Library

The craft of keeping people reading.

Claude-built visual studies of longform experiences, measured design tokens, editorial cadence, and a reference design system.

  • AI-assisted design
  • Computed styles
  • Design tokens
  • Visual research
  • Editorial UX
  • Design systems
  • Design
  • Research
Contribution & architecture

Translates visual observations into measured tokens, editorial cadence and a reference system for sustained reading experiences.

Claude-assisted visual analysis and editorial interface studies.

Design forensics · computed styles · editorial interfaces

Research

Longform UX Field Guide

Eight interfaces, under the microscope.

An independent reference collection with dated desktop/mobile evidence and a reusable comparison method.

  • No AI feature documented
  • Visual evidence
  • Mobile UX
  • Editorial UX
  • Accessibility
  • Interaction design
  • Research
  • Design
Contribution & architecture

Collects dated desktop and mobile evidence for eight reading interfaces and provides a reusable way to compare their behavior.

Independent desktop and mobile UX research, organized as a reusable field guide.

Visual evidence · accessibility · interaction research

Prototype

Annnimate Studio

Motion studies you can play with.

Original animation implementations with executable React examples and Next.js/Remix reference variants, derived from behavioral studies.

  • No AI feature documented
  • React
  • GSAP
  • Next.js
  • Remix
  • Animation
  • Interaction design
  • Frontend engineering
  • Design
  • Research
Contribution & architecture

Turns observed animation behavior into original executable examples, with React implementations and Next.js and Remix reference variants.

Original motion examples informed by observed interface behavior.

React · GSAP · Next.js · Remix

Prototype

RallyPro Brand System

A sport becomes a visual language.

A saved Open Design brand system with typography, color tokens, component kits, and dark/light variants.

  • No AI feature documented
  • Open Design
  • Design tokens
  • Brand design
  • Design systems
  • Component design
  • Design
Contribution & architecture

Connects typography, color tokens and component kits across dark and light variants in a saved brand-system export.

An Open Design brand-system export, separate from the Claude Design collection.

Brand design · design tokens · component kits

Saved Open Design export; not attributed to Claude Design.

Local build

Hospitality Platform Site

Make the service standard visible.

A multi-page hospitality site presenting staff, manager, and author workflows, product positioning, and service execution.

  • No AI feature documented
  • Next.js
  • Responsive design
  • Product storytelling
  • Hospitality technology
  • Design
  • Products
Contribution & architecture

Explains the staff, manager and author journeys through product positioning and responsive service-execution pages.

A local product-marketing website for hospitality service workflows.

Next.js · product storytelling · responsive design

Prototype

Figma Capture Research

From a live interface to design data.

A plugin workspace exploring interface discovery, DOM capture, and a shared schema for design tooling.

  • No AI feature documented
  • Figma
  • DOM capture
  • Design tooling
  • Schema design
  • Systems integration
  • Research
  • Design
Contribution & architecture

Explores how discovered interfaces and DOM structure could become shared design data through plugin and schema work.

A design-tooling prototype exploring DOM capture and shared interface schemas.

Figma · DOM capture · schema design

Research

Local Model Lab

Put the model on your own machine.

Model-location mapping, runtime experiments, and LM Studio/Ollama integration research for a local AI toolchain.

  • AI research
  • LM Studio
  • Ollama
  • Local inference
  • Systems integration
  • Technical research
  • Research
  • Agent systems
Contribution & architecture

Maps model locations and runtime behavior while exploring how LM Studio and Ollama fit into a machine-local AI toolchain.

Local inference and integration studies using LM Studio, Ollama and upstream models.

LM Studio · Ollama · local inference

Integration research, not authorship of the model runtimes.

Research

Agent Harness Lab

Make different agents work together.

Setup and verification studies around QM, OpenClaw, Odysseus, and shared agent tools.

  • AI research
  • QM
  • OpenClaw
  • Odysseus
  • Agent orchestration
  • Cross-harness workflows
  • Systems integration
  • Research
  • Agent systems
Contribution & architecture

Examines setup, workspace behavior and verification across QM, OpenClaw, Odysseus and shared agent tools.

Integration experiments across existing coding-agent harnesses.

Harness integration · workspaces · local verification

Experiments with upstream projects; upstream code is not presented as an original product.

Research

Agent Delivery Process

Turn agent activity into owned decisions.

A shared development-process study with durable findings, decisions, handoffs, and explicit release gates.

  • Agent tooling
  • Process design
  • Audit evidence
  • Delivery governance
  • Cross-harness workflows
  • Technical leadership
  • Research
  • Agent systems
Contribution & architecture

Makes findings, decisions, handoffs and release gates durable across harnesses so agent activity can be reviewed and owned by a team.

Process research for governing AI-assisted delivery through shared decisions and audit evidence.

Process design · audit evidence · cross-harness workflows

What could we
build together?

Bring your AI ambition, your product bet or your hardest workflow.
Let’s turn it into something people can use.

Email me
Hire me. Partner with me.

AI engineering.
Product leadership.
Fractional CTPO.

Connect on LinkedIn ↗Request my CV by email ↗

Protego

Have some magic, because we can