Mughni Qureshi

I’m an AI Engineer with 2 years of experience building production Generative AI systems—from RAG-based chatbots with vector search to real-time computer vision apps. I focus on practical deployment: scalable ML pipelines that serve 10K+ daily requests, with measurable user impact.

At Terafort, I built an end-to-end “QA control plane” for agentic mobile-release QA: analyzing an Android APK, AI-generating a risk-based test plan, executing it on a device, and emitting a PASS/RISKY/BLOCK decision with evidence. I also developed an 8-agent LLM pipeline using an MCP tool gateway with a full audit trail, plus a deterministic Eval Judge that gates release decisions with observability.

I’ve shipped high-scale AI product experiences, including an AI Photo Editor with enhancement pipelines and style transfer templates—delivered to 1M+ users on Google Play with a 4.0 rating. I also developed an AI-powered interior design app using computer vision and generative models, including image segmentation & object detection (YOLO/Mask R-CNN), and achieved 30% faster design previews by optimizing the inference pipeline.

My projects reflect my engineering range: a production-deployed domain expert RAG chatbot (LangChain, FAISS, FastAPI) and a serverless multi-user Discord weather briefing bot on AWS with OAuth2, event-driven architecture, and infrastructure as code. I’m driven by data-driven experimentation, collaborative teamwork, and Responsible AI—backed by my leadership as Data Science Lead at GDSC and being selected as a Certified GDSC Keynote Speaker.