Mehak batool User

At Letsremotify, I architect multi-agent AI platforms for eight enterprise customers, generating $200K ARR while serving 5K+ active users and processing 10M+ monthly requests with 99.9% uptime.

I build RAG systems that combine vector search, semantic search, and Neo4j knowledge graphs. My work improved retrieval accuracy from 72% to 95% and reduced ticket resolution time from 45 to 28 minutes.

I've fine-tuned Llama 2 and Mistral models, reduced hallucinations from 8% to 3.2%, and built production evaluation, observability, and monitoring systems with LangSmith and Datadog. I also lead a six-person engineering team and mentor junior engineers in production ML and AI-agent architecture.

Previously at Tkxel and IT Labs, I deployed NLP models, ML-powered applications, ETL workflows, and containerized services using Python, Docker, Kubernetes, React, Flask, and Django. I've focused on measurable improvements in accuracy, latency, deployment speed, reliability, and infrastructure cost.