All work

Real AI systems. Real architectures. Real tradeoffs.

A selection of products, platforms, and infrastructure I've built — from live SaaS with paying users to research microservices that taught me what not to do at scale.

● Beta / 2025–present / Founder & engineer

CiteAgent

A live AEO (AI Engine Optimization) and SEO platform that monitors how ChatGPT, Gemini, Perplexity, and Google AI cite your business — and ships fixes automatically.

Next.js 16 React 19 Turborepo tRPC Drizzle +7
Read case study
● Live / 2025–present / Founder & engineer

Akoota

A studio for small AI utilities — humanizers, prompt optimizers, email writers — that work instantly with no signup. Acts as the parent brand for the studio's products.

Next.js 16 Turborepo Payload CMS Clerk Postgres +3
Read case study
● Private / 2024–2025 / Founder & engineer

AI Tutor

A 5-service Python system that ingests books and the open web, stores them in a vector index, and adapts to a learner — built on hexagonal architecture, event-driven communication, and CQRS where it earns its keep.

Python 3.11 FastAPI RabbitMQ Postgres Qdrant +7
Read case study
● Private / 2025 / Founder & engineer

Pentagon

A standalone Go-To-Market platform that scrapes leads, scores them with an LLM, generates personalised cold emails, and runs the whole pipeline on a daily Inngest cron — for a sister product, Echo.

Next.js 15 React 19 TypeScript Drizzle ORM Supabase / Postgres +10
Read case study
● Private / 2025 / Founder & engineer

SocialAI

A pipeline that fingerprints a creator's face and style, then generates images and short videos that stay visually on-brand across thousands of outputs — a vector-first system on FastAPI, pgvector, and Inngest.

FastAPI Python 3.12 SQLModel asyncpg Postgres +6
Read case study

Working together

You have an AI feature in mind. I have shipped a few.

Whether you need a one-hour gut-check, an architecture audit, or a real pilot — we'll figure it out on a 15-minute intro call. No deck, no pitch.