Back to logs

Week 1

Week 1 Goals

  • [x] Set up my public engineering presence
  • [x] Start a flagship AI/system project
  • [x] Publish at least one deep technical write-up

What I’m Building

I’m starting with OpenSkill Atlas — a skill-intelligence platform that focuses on proof over resumes.

The goal is to infer real skills from code, projects, and activity instead of self-reported claims. In a world where everyone lists "Expert Python" on LinkedIn, I want to build a system that verifies it through semantic analysis of actual engineering output.

Technical Decisions Made

  • Public Documentation: Decided to document everything publicly. The pressure of visibility forces higher quality thought processes.
  • Modular Backend-First: Chose a modular approach. I am deliberately ignoring the frontend UI for now to ensure the API surface area is robust.
  • Architecture over Features: Prioritizing clarity of architecture over feature completeness.

What Broke / Was Confusing

  • Scope Creep: Defining scope without overengineering is harder than coding. I spent too much time debating which Vector DB to use before I even had data to store.
  • The MVP Line: Deciding what is "MVP" vs "nice to have." I had to cut the real-time analysis feature to ensure I ship v0.1.

What I Learned

"Visibility compounds faster than perfection."

Writing clarifies thinking. I actually realized my initial database schema was flawed while writing this log. Real engineers document trade-offs, not just success.

Next Week Focus

  1. Finalize OpenSkill Atlas architecture diagram.
  2. Start backend skeleton (likely FastAPI + PostgreSQL).
  3. Publish first deep technical article.