A full checklist for building AI/ML portfolio projects that actually unlock job offers
kmeanskaran · x · 2026-09-12
A practitioner shares a step-by-step route for AI/ML portfolio projects: high-level design, data pipelines, core models/agents, FastAPI, caching, Docker Compose, offline evals, model/data versioning, Prometheus/Grafana, CI/CD and response tracing. Key thesis: value now lies in optimizing cost, latency and UX, not building another OpenAI — most flashy projects can't even handle 100 requests/hour.
More from coding & agent
- Dev on Astra: a stubborn genius you shouldn't try to control, and how to use it — kevinkern · 2026-09-12
- Sentry ships Agent Plugin: one command wires AI coding assistants to fix production errors — zeeg · 2026-09-12
- steipete's agent stack: Ghostty over VS Code terminal, GPT-5 for plan reviews — charlieharris01 · 2026-09-12
- Claude Code 2.1.269 ships plugin eval command and prompt tokens jump 69% — ClaudeCodeLog · 2026-09-12
- Claude Code 2.1.269 ships 98 changes: plugin evals and Bash edit diffs — ClaudeCodeLog · 2026-09-12
- Claude Code 2.1.269 released with 98 CLI changes — ClaudeCodeLog · 2026-09-12