A collection of deep technical notes written as problem-first engineering documents. Each note follows the same structure — Problem → Constraints → Architecture → Trade-offs → Implementation → Failure Modes → Testing → Lessons — and every one is grounded in a real system I designed and shipped.
By EslaM-X.
| # | Note | Grounded in | Core idea |
|---|---|---|---|
| 01 | Policy-driven robotics | robot-sim-policy-lab | Separate what to do (policy) from how to move (controller) so behaviour is testable without physics |
| 02 | Sim-to-sim validation | robot-sim-policy-lab | One policy, three simulators; differences in physical engines are the test, not an excuse |
| 03 | Replay-safe payment systems | production-systems-lab + RoboPay | Idempotency keys, replay windows, and hash-chained audit as the minimum for financial-grade correctness |
| 04 | GitHub CI engineering | RoboPay CI | CI is a product with a contract: every job pins a commit, runs in isolation, and leaves proof |
| 05 | Secure API design | production-systems-lab | HMAC over JWT, token buckets, validation as a first-class layer |
| 06 | Execution proof protocols | proofx | From "trust me" to "verify it yourself" — claim → evidence → proof → verification; 102-case conformance, Go/WASM/Rust |
| 07 | Portable verifiable evidence | piproof | PEP/1 protocol — Proof Passports, Dispute Engine, cross-application proofs, zero dependencies |
Start with 01. The notes build on each other: robotics establishes the policy/controller split, 02 uses it for sim-to-sim honesty, 03 and 05 apply the same discipline to financial and API systems, and 04 explains how the CI that enforces all of it is built and audited.
These are intentionally personal engineering documents. Feedback and corrections are welcome via issues and pull requests.