Faraz Mubeen Haider
AI Engineer
Production RAG Systems · LLM Orchestration · Scalable Python Backends
Production Systems
Real systems serving real users. Not demos.
Coming soon...
Projects
HireMind AI
Multi-job resume matching with ATS scoring
Lesson: Multi-tenant RAG with user-specific document isolation
AI Research Assistant
Multi-source synthesis from ArXiv, Wikipedia, and web
Lesson: Agent orchestration with error recovery
LLM Document Summarizer
Multi-style summarization via Groq API
Lesson: Streaming responses and progressive enhancement
Hackathon Track Record
| IBM Granite | Winner | Production AI in 48h |
| DeepSeek DevAI | Finalist | Context-aware debugging |
| Replit x Cursor Aptimi | Finalist | Hybrid search +31% |
Insight: What hackathons taught me: I ship production-quality systems in 24-72 hours. This translates to rapid prototyping and calm execution under pressure. I've participated in 17+ international hackathons across various platforms.
Case Studies
Why Most RAG Systems Fail in Production
Most RAG demos break in production because they ignore evaluation, cost, and failure modes. Here's what I learned building one that didn't.
Building a 50K-Document RAG System on a $20/month EC2
Cost optimization isn't about cutting corners. It's about making the right architectural decisions early.
6 AI Concepts You Must Master for Production-Ready AI Systems
The gap between demo and production is wider than most engineers realize. These six concepts close it.
Technical Stack
AI/LLM
Vector/Search
Backend
Data
DevOps/Cloud
Observability
Testing
About
I'm an AI Engineer based in Pakistan, currently building production RAG systems at ClinicOps AI in Doha. I specialize in taking AI prototypes from notebook to production — handling latency, cost, evaluation, and failure modes most demos ignore.
Before ClinicOps, I was sole backend engineer on a 12-week AI SaaS MVP at 1337 Ventures, shipping with zero production incidents. I've won 17+ international hackathons.
I'm open to remote AI Engineering roles with US timezone overlap (6 PM - 2 AM PKT = 9 AM - 5 PM EST). Contractor-ready via Deel/Remote.com.
Education
- B.S. Software Engineering NUML (Jul 2026)
- Stanford Code in Place Section Leader (top 100/900+)
- Harvard CS50x Puzzle Day Perfect 9/9
Recognition
- UC Berkeley CALICO Top 12%
- Y Combinator Startup School
- Founder Institute Pakistan 2025
- 1337 Ventures Alpha Top 5%
Let's Build Something
Open to remote AI Engineering roles and consulting.
Remote-first · US timezone overlap · Contractor-ready
LinkedIn: linkedin.com/in/fm618 · GitHub: github.com/Faraz6180
Medium: medium.com/@farazmubeenhaider902