Amir User

I’m an impact-driven AI Engineer with a proven track record designing, fine-tuning, and deploying production-grade Generative AI, RAG architectures, and Computer Vision solutions. I’m deeply committed to open-source innovation and building human-centric AI tools that turn complex algorithms into practical utility.

I orchestrate multi-agent systems, optimize LLM pipelines, and deploy scalable architectures across major cloud environments. I focus on reliable integrations and performance—so AI works in real-world workflows, not just prototypes.

At Pixelpk, I design, develop, and deploy end-to-end Generative AI products using Large Language Models, multi-agent frameworks, and advanced RAG architectures. I fine-tune open-source models (including DeepSeek, YOLO, and MedGemma) for strict domain constraints, and I architect robust data ingestion and preprocessing ETL pipelines for real-time speech and video AI streaming integrations.

Previously at Ekkel AI, I collaborated on training, evaluating, and optimizing production-grade ML and NLP models for commercial clients. I also built microservices with FastAPI and Flask, and engineered web scraping + ingestion pipelines to clean and structure proprietary text datasets—then applied these strengths across projects like MedNotes.ai (local, privacy-first LLMs for HIPAA compliance), conversational RAG chatbots, and NLP evaluation systems.