AI Consultant · AI Engineer
I build AI agents and automations that still run in month six.
4 years as a data scientist and ML engineer before this. That is the difference between an automation that demos well and one your team can depend on.
The problem
The work that quietly eats your week
The report someone rebuilds every Monday
Hours a month spent copying numbers between tools into the same spreadsheet.
The inbox nobody has time to triage
Enquiries sit unanswered, leads go cold, replies get written from scratch every time.
The documents nobody can search
Contracts, invoices and internal knowledge locked in PDFs and folders.
Why most AI automations break
Most break in month two. The demo works, then a rate limit hits, an API changes, a bad output goes unchecked, and the team quietly goes back to doing it by hand.
That is not an AI problem. It is an engineering problem.
How I build systems that last →Selected builds
Real systems, published in full
Research Papers Intelligence
An agentic pipeline that ingests thousands of academic papers, builds a retrieval-augmented knowledge base, and answers complex research queries with cited sources.
Local voice transcription, a free WhisperFlow alternative
A dictation tool built with Claude that runs entirely on my machine. No subscription, no audio leaving the laptop.
Local LeetCode-style interview trainer
A practice environment for SQL, Python, ML, AI and system design questions, with AI feedback, running fully locally.
