Moiz Imran

I am a Computer Science graduate from the Lahore University of Management Sciences (LUMS) specializing in AI engineering. My primary focus is building complete, production ready AI applications and scalable software architectures.
Below are some the projects I have worked on:

• Agentic AI & RAG: I engineered Nova Desk, an AI concierge for medical spa operations. I designed a workflow utilizing multiple agents via LangGraph and LangChain, integrated a Gemini and Pinecone RAG pipeline, and ensured reliability with PostgreSQL backed session checkpointing. The system runs on a React 19 frontend with a strictly typed FastAPI and SQLModel backend deployed across Render, Vercel, and Neon.

• Full-Stack B2B Systems: I developed a complete inventory and sales platform for ARKAD Mines and Minerals using the MERN stack. I built FastAPI microservices for demand forecasting and a Gradient Boosting machine learning model for quotation suggestions, alongside a CRM module and QR stock tracking. For the infrastructure, I utilized containerization using Docker and Kubernetes, configured automated CI/CD pipelines, and deployed the distributed system across Vercel for the frontend and Render for the backend services.

• Computer Vision & Deep Learning: I have a strong foundation in implementing, analyzing, and training advanced architectures in PyTorch. My experience includes Vision Transformers, Generative Adversarial Networks, Variational Autoencoders, and latent diffusion models.