5 projects
baby-llm
A developer-first PyTorch engine for explicit, modular LLM training. Zero magic.
skill-scoreboard
Benchmark Claude Code skills against a no-skill baseline, detect token waste and skill bypass, and rank a whole skill library KEEP/FIX/MERGE/RETIRE.
tunepac
YAML-configurable LLM fine-tuning harness built on TRL, PEFT, and Accelerate.
bts-pivot
Auto-fitted pivot tables that toggle to histograms and back, with slicing, clustering, Markov chains, paging for big data, an MCP server for LLM agents and Jupyter menus. For cyber logs or any tabular data.
supersafe
Unified evaluation framework for agentic LLMs: security red-teaming, tool-safety guardrails and latency/cost profiling across models.
