AI Backend Engineer | RAG Systems, LLM Integration & Voice AI | FastAPI, LangChain, PostgreSQL — Production-ready AI systems that scale
I'm a backend and AI engineer based in Lahore, Pakistan, specializing in building and stabilizing production-grade AI systems. My core stack: FastAPI, LangChain, ChromaDB, PostgreSQL, Celery, Redis, and LLM integrations with Claude, GPT-4o, and Retell AI.
What I do:
Design and build RAG (Retrieval-Augmented Generation) pipelines — document ingestion, chunking, vector search, and retrieval tuning
Develop multi-agent AI architectures for complex automation workflows
Build voice AI automation systems using Retell AI and LiveKit for real-time conversational agents
Debug and stabilize production AI backends — dependency conflicts, deployment failures, database migrations, and scaling issues on AWS EC2 and DigitalOcean
Set up AI reliability monitoring with Langfuse, Prometheus, and Grafana
Recent work includes building a Workday HR voice automation system (Retell AI + GPT-4o + Kubernetes) for enterprise use, developing an AI-powered audit platform with a multi-agent RAG architecture, and stabilizing several production AI systems handling live client traffic.
What sets me apart: I have a business background (BBA in Marketing) alongside my technical engineering work. This means I don't just write code — I understand the commercial context behind what I'm building, which is useful for client-facing MVPs, demos, and products that need to work in the real world, not just in a sandbox.
If your AI backend is broken, stuck in prototype mode, or you need a RAG/LLM system built from scratch, I can help you get there.
Work Terms
Communication: I respond within 12-24 hours, typically faster during active projects. I'm based in Pakistan Standard Time (UTC+5) and flexible with scheduling for meetings.
Process: For new projects, I start with a discovery discussion or written brief to understand requirements, followed by a scoped proposal with timeline and milestones. For ongoing or hourly work, I provide regular progress updates and stay responsive through the platform.
Payments: I work with Guru's SafePay for milestone-based or hourly billing. For fixed-price projects, I typically request 30-50% upfront depending on scope, with the remainder tied to milestones or delivery.
Revisions: Reasonable revisions are included as part of scoped work. Major scope changes are quoted separately to keep timelines realistic.
I'm open to both long-term collaborations and one-off projects, including retainer arrangements for ongoing AI system maintenance or development.