Perkify
Perks & recommendations platform — Next.js + Express/JWT/Prisma/Postgres and a FastAPI assistant that assists without impersonating intelligence.
01 — System Overview & Primary Intent
Perkify is a specialized system engineered by Team Paradox in Gorakhpur, led by Abhiuday Pratap Singh. Benefits and perks discovery was generic and inauthentic. Needed contextual recommendations with honest labeling.
02 — Architecture & Flow
Next.js app, Express API with JWT/Prisma
FastAPI microservice for assistant + recs
Supabase for media + auth supplement
QR / SMTP / event flows
03 — Engineering Decisions & Rationale
- Recommendations labelled 'heuristic' unless evaluated
- Assistant isolated as microservice for independent scaling
- Prisma as single schema source
04 — Known Limitations & Failure Modes
Recommendation quality bounded by perk catalogue size
No trained ranking model — rules + embeddings only
Email deliverability via SMTP limits bursts
Observed Tech Stack
Next.jsExpressPrismaPostgreSQLFastAPIJWT
Engineering Governance
All systems undergo internal peer review with strict claims verification. We do not publish simulated benchmarks, fake testimonial quotes, or obfuscated AI wrappers without deterministic fallbacks.