CS undergrad building practical AI/ML systems — RAG pipelines, LLM apps, and applied ML from scratch
I'm a Computer Science student at FAST-NUCES (Class of 2027), currently focused on AI/ML — I did a summer internship at AI Genmat working on applied ML workflows, and I build full end-to-end projects outside of coursework to actually ship things rather than just follow tutorials.
- 🔭 Currently building LLM-powered tools (RAG pipelines, structured extraction, local-first AI apps)
- 🌱 Learning more about production ML systems, vector databases, and agent workflows
- 💬 Ask me about Python, LangChain, RAG, or classic ML (regression, SVM, clustering)
- 📫 Reach me at umarhussain.ml@gmail.com
|
🎥 AI Video Assistant — RAG Chat Turns any YouTube video or local file into a searchable, chattable knowledge base. Local Whisper transcription → structured extraction (summary, action items, decisions) via Mistral → RAG chat over the transcript using LangChain + ChromaDB.
|
🧠 Reading Comprehension & Quiz Generation Two-pipeline system trained on the RACE dataset (~100K questions): one model handles answer verification and question generation, the other generates distractors and hints via cosine-similarity ranking. Wrapped in an interactive Streamlit quiz UI.
|
