Engineering Decisions
Why I chose what I chose
Chose PostgreSQL over MongoDB for AuraStream
2026-03Reason:ACID compliance needed for billing & user sessions. Relational schema fits the video metadata model.
Tradeoff:More schema rigidity, required migrations for schema changes.
Result:Zero data consistency bugs in 3+ months of production.
Pure CSS injection over Shadow DOM for ClearFeed
2026-02Reason:Shadow DOM creates isolation issues with YouTube's existing styles. CSS injection is simpler and 0ms overhead.
Tradeoff:Risk of YouTube style changes breaking the extension.
Result:No breakage in 3 months. MutationObserver handles DOM shifts gracefully.
Supabase + PostgreSQL for TrueKhoj instead of Firebase
2026-05Reason:Need relational queries for marketplace (products ↔ sellers ↔ categories). Firebase's NoSQL would require denormalization.
Tradeoff:Smaller community, fewer tutorials compared to Firebase.
Result:Sub-100ms query response times with proper indexing. Full SQL power.
Next.js SSR over client-side React for Invoice Maker
2025-11Reason:SEO for the landing page + faster initial load for PDF generation UI.
Tradeoff:Slightly more complex deployment config.
Result:30% load time reduction. Mobile usage up 40%.
System Log
Aspirations & Architectural Ambitions
01> Engineer a high-precision engine to detect AI-generated imagery and restore digital trust.
[IN FLIGHT]
02> Explore the neon-lit streets of Tokyo — a pilgrimage for design and discipline.
[QUEUED]
03> Architect a system that survives its first million users without a rewrite.
[QUEUED]
04> Design a UI so seamless it feels like part of the OS.
[IN FLIGHT]
05> Launch a production-grade video engine.
[REFINED]
06> Engineer a productivity tool for the elite.
[REFINED]
07> Build a tool that lasts a decade in a developer's daily workflow.
[QUEUED]
08> Earn $100,000 by the end of this year.
[QUEUED]
Progress0%
[LOG_END]: End of buffer. Syncing with reality...
Heartbeat stable.