Curiosity leads my way through life.
AI / LLM engineer building agents, RAG systems, and developer tooling, mostly in Python.
I spend most of my time on applied LLMs: agentic systems, retrieval-augmented generation, and tooling that makes models actually useful. A few things I'm proud of:
claude-audit, a command-line code auditor: fast static checks, plus an optional audit where Claude investigates the code with read-only tools.llm-ocr, which turns PDFs and images into clean Markdown with vision LLMs: Gemini, OpenAI, Claude, or a local model.async-limiter, an asyncio rate limiter that enforces a concurrency limit and a requests-per-period limit together.titans-unofficial, a from-scratch, unofficial PyTorch implementation of Titans ("Learning to Memorize at Test Time").DocuSenseAI, a multimodal RAG for pulling answers out of messy documents.
In between, there's usually a bot or a parser being built somewhere, and the occasional upstream contribution, like Ollama support for OpenVoiceChat.
Languages
AI / LLMs
Speech & Audio
Full-Stack & Data
DevOps & Tools
Always down to talk LLMs, agents, or a good side project.