I develop LLM apps, AI agents & production ML pipelines by turning data into informed decisions and creating sustainable Human-AI systems
| Project | Description |
|---|---|
| llm-cassette | Record and replay OpenAI and Anthropic calls plus tool calls in pytest, with a diff command.pip install llm-cassette |
| ServiceOS | AI-powered operating system for service businesses. |
| BillSplitPro | A minimal bill-splitting app for splitting by item, percentage, or evenly, with optional service charge, VAT, and tip. |
| Climate Change RAG Assistant | A retrieval-augmented generation assistant that answers climate change questions from research papers, using vector embeddings with ChromaDB, Google Gemini, and a Streamlit app. |
- pandas-dev/pandas-stubs #2012: pass --preset strict to the installed-stubs pyrefly check (merged)
- docling-project/docling-pipelines #355: bring vault_initializer.py to 100% coverage and fix its log dashes (merged)
- docling-project/docling-pipelines #430: support file events in the S3 and IBM COS ingest provider and load a single object key (open)
- mozilla-ai/otari #2525: add an Azure OpenAI provider guide for the resource endpoint, deployment names, and the v1 route the provider calls (open)
ML frameworks
LLM / agents
Vector DBs
AI automation
AI coding tools
Languages
Web
Backend and APIs
Cloud platforms
Data and ETL
Visualisation / apps
MLOps and DevOps
Tools



