Product engineering
I take ideas from a blank page to production. I care about the data model, the API and the last 10% of the UI, and about whether anyone uses the result.
Software engineer taking ideas from a blank page to production — web, mobile and, lately, agentic AI. I care about the whole thing: the data model, the API, the pixels, and whether anyone actually uses it.
I'm happiest owning a product slice end to end — talking to users, shaping the data model, building the API and polishing the last 10% of the UI.
I take ideas from a blank page to production. I care about the data model, the API and the last 10% of the UI, and about whether anyone uses the result.
React Native and Expo apps from first screen to store, with the push and release pipeline that keeps them alive.
Spring Boot and Node services, event-driven pipelines and data models that won't need rewriting in six months.
Agents and automations that plan, call tools and hand off to humans — designed to be measured, observed and trusted inside a real product.
Products I founded or co-founded — designed, built and shipped to the web, the App Store and Google Play.
Building full-stack AI products — an expense tracker with an embedded chat assistant, RAG pipelines on the OpenAI API and Pinecone, and multi-agent orchestration that plans tasks and hands off to humans.
Built a chat-driven knowledge-graph interface with the Vercel AI SDK, React, Redux-Saga and TypeScript, backed by Node.js APIs, Azure Functions and a full RAG architecture deployed on Azure.
Built the SaaS marketplace from scratch and shipped it across web, iOS and Android — a scalable multi-tenant Node.js backend, React/TypeScript dashboards and React Native + Expo mobile apps.
Owned the Customer Data Platform UI in React.js, Redux-Saga and TypeScript, and built Node.js onboarding automation on AWS (S3, Lambda, API Gateway).
Proven tools for the foundations, so the novelty budget goes where users can feel it.
A write-up of the build — the decisions, what broke, and what I'd do again.
How an AI agent actually works: call the Messages API, check stop_reason, run the tool, append the result, repeat until the model is done.
A walkthrough of the stack, the folder structure and how the Markdown blog works — so you can make it your own.
Always up for a good product conversation — a role, a collaboration, or just comparing notes on shipping.