Curriculum vitae
Chukwuemelie “Chuk” Uyammadu
AI Systems & Data Infrastructure Engineer
Summary
I build AI-assisted systems and the infrastructure that runs them. My work covers backend services, local infrastructure, and business operations. I use human review and read-only data access where risk matters. Each deployment includes a written handoff and rollback path.
Selected work
Orion — Read-only AI Reporting Prototype
Python/FastAPI prototype that turns business questions into structured BI responses. It uses fictional CSV data and read-only connectors. Guardrails block destructive operations. Deterministic evaluations and decision logs make the results reviewable.
ForgeWorks / ForgeTable — Local builder community platform
Rust, Axum, and SQLite prototype with public pages and gated operator tools. One API supplies build metadata and homepage counts. The platform also has RSVP intake and a prelaunch public-page check.
Peppino’s — Restaurant technology operations
Supports a restaurant website, local SEO, POS operations, cameras, and network troubleshooting. The work prioritizes uptime and written handoff. Internal details and identifying data stay private.
Crenshaw — Local agent + ntfy workflow
Local agent workflow that posts short ntfy checkpoints during build, edit, and audit cycles on this site and related projects. Used to keep human-in-the-loop review fast without exposing secrets or hardcoding tokens.
Raspberry Pi homelab / local services lab
Repeatable Raspberry Pi infrastructure for local services and dashboards. Uses Linux, Docker, systemd, Tailscale, and Cloudflare Tunnel. The lab is a reference for small-business systems that run on owner-controlled hardware.
dev.uyammadu.com — Static portfolio & services site
Static HTML and SASS site deployed on Cloudflare Pages. A Pages Function handles the contact form with a honeypot, length cap, and Resend integration. The site uses clean routes and has no tracking scripts or JavaScript framework.
Working principles
- Scope the business problem in writing before picking a tool.
- Use read-only workflows where risk matters. Make changes easy to inspect and reverse.
- Keep public claims bounded to what is actually built or demonstrable.
- Document the change and its rollback steps.
- Give AI systems typed plans, guardrails, and human review.
Live version: dev.uyammadu.com/cv · References, private case-study walkthroughs, and reviewer access available on request.