Mohammed Yaseen

I’m currently pursuing a Master of Engineering in Artificial Intelligence and Machine Learning at Osmania University, building my work around practical, measurable AI systems. My focus is turning models into reliable applications—whether that’s improving code generation quality or answering domain-specific questions with grounded context.

At InI Labs, I worked as a Python Developer Intern with the Data Curation Team, where I cleaned, validated, and structured raw data. I developed Python scripts to convert unstructured text into organized HTML-tagged data, and I optimized existing Python code to reduce execution time and improve processing efficiency.

My projects reflect how I approach end-to-end delivery: research-grade LLM tuning plus production-ready pipelines. I fine-tuned Qwen 2.5 Coder 32B with QLoRA (Unsloth, PyTorch, PEFT, TRL, Modal) and improved pass@1 by 3.35% on BigCodeBench-Hard, while enhancing docstrings, modular structure, and error handling. I also built LawRAG with LangChain, SentenceTransformers, ChromaDB, and Gemini, including evaluation with RAGAS/BERTScore and Docker-based deployment. Finally, I created and deployed Imagio, an AI Stories Generator mobile app with FastAPI, Flutter, Firebase, and Google Cloud Run—reaching 1700+ organic downloads and 500 monthly active users.