Abdullah
ai engineer · agents & evals
I build agentic systems, LLM orchestration, evaluation, and the guardrails that make them trustworthy. Eight years shipping production software, mostly TypeScript and React.
experience
- NextfinitiSenior Software Engineer / Technical Consultant2022 – present · Remote
projects
- Sidekick ↗Self-verifying agent. A worker drafts, a second model judges against explicit criteria, and the worker revises until it passes — with a 10-case deterministic eval harness.
- Baton ↗Shared memory for AI agents. An MCP server and CLI that carry task, decisions, and next steps across Claude, Cursor, and Codex handoffs.
writing
- Free is never freeSeptember 5, 2022
- Writing Perfect Code?July 14, 2021