I taught myself AI through Stanford and DeepLearning.AI's Machine Learning and Deep Learning Specializations on Coursera. I then built and deployed a production crypto analytics platform from scratch, including an LLM sentiment pipeline with a fallback model hierarchy and an LSTM model for price forecasting. I'm currently extending it with a LangChain multi-agent architecture and MLflow for experiment tracking.
Houdini work sits at the intersection of art and engineering. Building a convincing fire, fluid, or destruction effect means turning physics into math and then into procedural systems written in VEX and Python. Each shot is a complex dataset of particle counts, force fields, and collision geometry that must read as physically real. Few people work comfortably across physics, math, and code at once, and that combination let me build custom solutions when standard workflows fell short.