CanadaApollo6/agent-training
Playground for me to try different things with RL and post-training
Senior full-stack engineer building AI tools, production ML systems, and model-evaluation experiments. React/TypeScript + Python.
Playground for me to try different things with RL and post-training
An open source agentic harness for video editing
Config files for my GitHub profile.
Scalable toolkit for efficient model reinforcement
Record meetings on Omarchy: mic and computer audio as two tracks, transcribed on your own machine, with speakers, chapters and a player.
Computational PSC research using public genetics, AlphaGenome, and evidence tracking.
An unofficial open-source client for ESPN's various API endpoints. Written in TypeScript for full type safety and IDE hints.
My KCDC 2022 AI & Machine Learning Workshop materials
An interactive lab for a Snowflake Cortex workshop given to Smart Data clients
Deductive-ground-truth evaluation of forecasting methods — what learners recover vs. what's only reachable by rule, foundation models included.
GRPO + verifiable rewards — teaching a small LM to reason over structured data on one GPU.
Thesis synthesis of previous 3 papers
Expand upon work in paper 1 across domain sets
Production COVID-19 diagnostic system. Ensemble CNN-LSTM networks deployed via TensorFlow.js for real-time inference.
React dashboard for COVID-19 diagnostic system. Real-time visualization of PCR curve analysis and ML inference results.
This is a simple gradient and background generator, allowing users to select any 2 colors via a color wheel, see the resulting gradient, and the CSS required to generate the given background.