Software Engineering
│
▼
Backend · APIs · Data · Reliability
│
▼
AI-Native Systems · Agents · Automation
│
▼
Product Engineering
│
▼
Useful Digital Products
│
▼
Measure → Learn → Improve → Ship again
I use GitHub as a working engineering portfolio: real systems, real experiments, real failures, and progressively stronger products.
Portfolio system map: SYSTEM-MAP.md — canonical map of projects, shared capabilities, real dependencies, gates, and the future operating surface for long-running agents/Dots.
My direction is deliberately broader than one niche:
software engineering + AI-native systems + developer tools + technical product building
|
APIs · PostgreSQL · Testing · Git · CI/CD · Backend architecture Goal: build maintainable systems with clear contracts and production-oriented behavior. |
RAG · evaluation · observability · agentic workflows · automation Goal: treat AI as part of a real software system, not as an isolated prompt. |
Distribution · retention · pricing · packaging · product economics Goal: turn engineering capability into products people can understand and use. |
TLA+ · state machines · invariants · concurrency · failure modeling
I use formal methods as an engineering discipline for understanding complex systems, especially distributed state, concurrency, and failure behavior — not as the identity of the portfolio.
|
CI / Developer Tools Experiments around failure analysis, retry behavior, CI reliability, and practical engineering automation.
|
Databases / Reliability Measures regressions around PostgreSQL changes and tests stronger ways to explain what actually changed.
|
|
Product / Analytics Explores whether reported conversion reflects the real customer journey and whether measurement can support product decisions.
|
AI + Software Engineering Uses AI in code modernization while keeping testing and engineering verification inside the implementation loop.
|
|
Cloud / Reliability Detects failures in authentication-email delivery flows.
|
AI / Deployment Experiments around shipping and operating AI-enabled software.
|
Build real things. Learn from the implementation. Measure what happens. Improve the product. Ship again.
I prefer projects that force useful learning: APIs that must work, workflows that must survive failure, databases that must preserve state, AI features that must operate inside software, and products that must produce understandable outcomes.