I'm a software engineer who believes tools don't matter anymore.
What matters is understanding fundamentals deeply enough to make the right tradeoffs — and being fast enough with AI tooling to actually ship. I think in two modes: engineering (how does this scale, what breaks first, what's the tradeoff) and product (does this solve a real problem, would a real person use this, what's the shortest path to value). Both matter. Most engineers only have one.
I'm 21, studying Computer Engineering at Shahid Beheshti University — where I built my foundation in things that don't expire: OS internals, database theory, data structures & algorithms, software engineering principles, requirements engineering. The frameworks change. The fundamentals don't.
| Project | Description | Stack |
|---|---|---|
| CoachUpgrade | SaaS marketplace for Iranian fitness coaches — zero to production in 2 days, solo | Django · HTMX · PostgreSQL · Redis · Docker |
| ApplyBuddy | Multi-agent AI platform for university applications — RAG pipelines, 5 specialized agents, hundreds of active users | LangGraph · LangChain · Qdrant · Django · FastAPI |
| Age of AI | Production AI chatbot infrastructure — built the backend from scratch | Django · FastAPI · Redis · LangGraph · Claude API |
- 85% of requests served under 1 second — Redis multi-layer caching
- 40% latency reduction — async concurrent retrieval with
asyncio.gather() - 10,000 concurrent users — load tested with k6, p95 under 1 second
- 2 days — CoachUpgrade, zero to production, sole engineer
Core
AI
Data
Infra
Frontend (when needed)
- Building and validating CoachUpgrade with real Iranian fitness coaches
- Looking for a backend or AI engineering role at a company with real scaling problems
- Reading: OS concurrency internals, Go runtime, transformer architecture