Game Dev turned Data Choom turned Data Scientist turned Dragonborn turned ML Engineer.
Voice Agents. Automation Agents. Game Agents.
Breaking things to understand them, then making them work in production.
- Production voice AI that doesn't collapse under load (Pipecat, WebRTC, AssemblyAI)
- Systems programming in Rust because someone has to do memory management
- Mechanistic interpretability—opening the black box with a crowbar
- Procedural generation experiments (worlds build themselves)
Languages: Python for speed, C++ for performance, Rust for when I hate myself
ML/AI: PyTorch, TensorFlow, scikit-learn, LLM wrangling
Infrastructure: Docker containers all the way down, AWS, GCP, Spark/Kafka
Voice: Real-time streaming, WebRTC, making very expensive large predictive models talk
Data: SQL, MongoDB, Voldemort,Cassandra and pretty much all the VectorDBs
Game Engines: Unreal, Unity (for when I am reminded how bad the RAM market is)
I like data. I like machines. I like learning.
Prefer understanding how AI works over just calling APIs.