Muhammad Saim

I’m a CS student at FAST NUCES and the founder of Vexilot.dev, a production AI platform serving 2,100+ users. I build full-stack AI products at the intersection of engineering and product thinking, and I’ve shipped RAG systems, real-time execution pipelines, and gamification engines—largely solo.

At Vexilot.dev, I designed and deployed a competitive programming platform serving 1200+ active users within two weeks, combining real-time code execution, AI-powered tutoring, and a gamification engine. I built a RAG-powered tutoring system with LangChain + Gemini that reduced query response time 90%, implemented a PostgreSQL (Supabase) schema with 20+ tables and Row-Level Security, and engineered a secure execution pipeline integrating Judge0 with Piston fallback (C++, Python, Java) with sandboxing and 99.5% uptime. I also focus on platform integrity—rate limiting with Redis/Upstash, XSS/SQL injection prevention, security logging, and AI-based plagiarism detection—while delivering production reliability through Docker, CI/CD, and zero-downtime deployments to Vercel/Cloudflare.