20 field-tested AI/ML rules: never let coding agents run commands themselves
kmeanskaran · x · 2026-11-19
An AI/ML practitioner shares hard-won lessons: - Don't chase every trend; make every project end-to-end. - Learn high-level ML system design; explain projects via business, data pipeline, model/agent, eval, deployment, and numerical results. - Never let coding agents run commands themselves — have them produce commands and run them yourself; always check vibe-coded results against expectations. - Companies care about impact, not your stack; interviewers want knowledge, not buzzwords. - MLOps is mandatory — no specialized ML role remains; know Airflow, MLflow, Linux, Docker, CI/CD, Terraform, Redis; Bash/YAML/Markdown/JSON are 90% of engineering. - Automate low-level coding with LLMs but keep a human in the loop. - GraphRAG is the most important thing to learn; ETL design is high-leverage. - Firms treat AI, ML, MLOps, and data as one department — learn all of it.
More from coding & agent
- Open-source t3boat runs T3 Code agents on cloud machines that sleep when idle — gethackteam · 2026-11-19
- CS Professor Lemire Slams AI Slop in GitHub Issues: 'Answer Me, Don't Throw Slop at Me' — lemire · 2026-11-19
- Univer hits #1 on GitHub Trending: open-source AI office SDK with 29.9k stars — alifcoder · 2026-11-19
- Running a Personal AI Agent on One RTX 4090: Open-Source Linger Plus PolyStrata, 5x Faster Than llama.cpp — lxfater · 2026-11-19
- User builds tower defense game in 20 minutes with Gemini 4 Neon, using only 2% of quota — DeryaTR_ · 2026-11-19
- ProtoInputMultiPlex: an MCP server giving AI agents their own virtual mouse and keyboard on Windows — Kenzillla · 2026-11-19