A focused list of work with public proof: repositories, write-ups, and live links.
A lightweight evaluation workbench for comparing open-source and frontier AI assistants before enterprise deployment. Surfaces real differences between model backends with structured CSV results, one-page PDF reports, and auditable risk findings across hallucination, bias, jailbreak resistance, and cost.
Python · Chainlit · LLM evaluation · Guardrails · Modal
A public learning repo structured as 12 stations across the LLM engineering stack: SFT, structured output, evals, vLLM serving, guardrails, RAG, and observability. It is the upskill pipeline feeding V2 of the AI Risk Evaluation Workbench.
SFT · LLM evaluation · vLLM · Guardrails · RAG
An OpenClaw plugin for fetching and analyzing GitHub issues, built from the hackathon work I wrote about. Reduces repetitive project work with AI-assisted tooling.
GitHub workflow automation · LLM tooling · OpenClaw