OBS Power: A Power brokage service for B2B

2/28/2026• Dulranga Dhawanitha

1. Problem Context (Why This Exists)

  • What OBS Power actually does (1–2 sentences, no marketing fluff)
  • The exact problem you were solving
  • Why existing solutions were insufficient (be specific, not vague claims)

👉 If this section is weak, the whole post is pointless.

2. High-Level System Overview

  • Core entities (users, offers, traders, teams, etc.)
  • How the system flows end-to-end (lead → processing → output)
  • A simple architecture diagram (don’t skip this)

3. Tech Stack (With Justification)

Break it down properly:

  • Frontend: Next.js (why? SSR? routing? SEO?)
  • Backend: Express.js (why not serverless fully?)
  • Database: MongoDB (why this over SQL?)
  • Infra: AWS (which services and why)

👉 If your answer is “because I’m familiar with it,” that’s weak. Say tradeoffs.

4. Core Features Breakdown

Don’t list features—explain mechanics.

4.1 Offer Request System

  • How requests are created
  • How traders are assigned
  • Data relationships

4.2 Trader + Team Structure

  • Why teams were introduced
  • Team leader logic
  • Assignment flow changes

4.3 Offer Generation Flow

  • What happens after submission
  • How multiple offers are generated
  • Any automation logic

5. Key Technical Challenges (This is the real value)

Pick 3–5 actual problems you struggled with:

Example:

  • Scaling offer assignment logic
  • Maintaining consistency across async flows
  • CDN caching vs dynamic data issues
  • Redirection system (Lambda@Edge + DB lookup)

For each:

  • What broke
  • Why it broke
  • What naive solution failed
  • Final solution (with reasoning)

👉 This section separates a serious engineer from someone just documenting CRUD.

6. Architecture Decisions (Tradeoffs)

  • Why not microservices?
  • Why not serverless everywhere?
  • Why MongoDB despite relational data?
  • Where you knowingly accepted technical debt

👉 If you pretend everything is perfect, it reads like nonsense.

7. Performance & Optimization

  • Caching strategy (CDN, endpoints)
  • DB query optimizations
  • Handling high request volume
  • Any measurable improvements

8. Security & Reliability

  • Auth approach (NextAuth, tokens, etc.)
  • Data validation
  • Preventing abuse (rate limits, etc.)
  • Failure handling (retries, fallbacks)

9. Deployment & DevOps

  • Docker setup (multi-service)
  • CI/CD (GitHub Actions → ECR → deploy)
  • Environment management
  • Any production issues you hit

10. Lessons Learned (No sugarcoating)

  • What you’d do differently
  • What turned out harder than expected
  • Bad decisions you made early

👉 If this section is honest, it builds credibility. If not, it’s just self-promotion.

11. Future Improvements

  • Features you plan to add
  • Architectural changes needed
  • Scaling concerns

12. Conclusion

  • Short recap
  • Who this system is actually useful for
  • Optional: link to repo/demo

Optional (but high value)

  • Screenshots of system UI
  • Architecture diagram
  • Code snippets for critical parts (not random code)

If you follow this structure strictly, you’ll produce something worth reading.

If you skip the “challenges” and “tradeoffs” sections, you’re just writing marketing disguised as engineering.