I build

Autonomous AI Agents

Automated Workflows

Full-Stack Web Apps

Landing Pages That Convert

Custom AI Chatbots

Business AI Assistants

Stop wasting time on repetitive tasks. Let intelligent automation run your backend while you focus on scaling.

Start Your Project
By the Numbers

Track Record

What I've shipped, earned, and organised so far.

0+

Projects Delivered

Shipped and running in production

0+

AI Systems Built

Agents, RAG pipelines, and LLM apps

0+

Certifications

AI, ML, and cloud coursework

0+

Events Led

Conferences, workshops, and meetups

Services

What I Do

The kinds of systems I design, build, and put into production.

Web Applications

Full-stack apps where the schema and the interface get designed together: typed from the database through to the component, server-rendered, with row-level security so one account cannot read another's rows. Includes the unglamorous parts that decide whether it survives real users, like auth, migrations, and running the whole interface in a second language.

Stack

Next.jsTypeScriptSupabaseTailwind CSS

Computer Vision

Detection, recognition, and matching on real photographs rather than benchmark sets. That means the blurry frames, the faces turned half away from the camera, and the near-duplicates a naive similarity threshold will happily merge into one person. I tune the threshold against your data, because the right cutoff for a wedding album is not the right cutoff for a door lock.

Stack

RetinaFaceArcFaceOpenCVCUDA

Data Pipelines

Scheduled jobs that pull from APIs and databases, reconcile the messy parts, and land clean rows somewhere you can actually query and chart.

Stack

PythonSQLSupabaseREST APIs

Workflow Automation

Connecting the tools a business already runs on, so the copy-paste steps between them stop being someone's job. Usually n8n where it fits, custom integrations where it does not.

Stack

n8nREST APIsWebhooksPython

RAG & Retrieval Pipelines

Chunking, embedding, and retrieval over your own documents so answers point at something real. Including the part most demos skip: what the system does when nothing relevant comes back.

Stack

LangChainEmbeddingsVector SearchPostgreSQL

Model Training & Fine-Tuning

Training and fine-tuning on your data, plus the evaluation harness that tells you whether the new checkpoint is genuinely better than the one already in production.

Stack

PyTorchTensorFlowScikit-learnXGBoost

Backends & APIs

The service layer underneath the model: streaming endpoints for anything that updates live, auth and middleware, and enough queueing that one slow inference call does not block every other request.

Stack

FastAPIWebSocketsAsync PythonREST

Multi-Agent Systems

An agent that plans its own steps, calls tools, and knows when it is done. I build the graph, the tool definitions, and the fallbacks for when a step fails or a model returns nonsense.

Stack

LangGraphLangChainTool CallingPython
Swipe to navigate
Technologies

Tech Stack

Tools and technologies I use to build intelligent systems

Next.js
React
TypeScript
Tailwind CSS
Python
PyTorch
TensorFlow
LangChain
LangGraph
Hugging Face
OpenAI
Anthropic Claude
FastAPI
Node.js
PostgreSQL
Supabase
Redis
Docker
Git
Linux
AWS
Vercel
n8n
Zapier
Express
Jupyter
Get Started

Tell me about your project

Describe what you're building and where it's slow. I'll tell you what I'd automate first.

Form Progress0/7

What to expect

  • A free 30-minute call
  • A reply within 24 hours
  • A written scope, with no commitment
  • Pricing before any work starts

Include country code (e.g. +1, +44).