I spent three years as a Software Engineer at Cognizant maintaining and modernizing enterprise .NET applications on Azure - building CI/CD pipelines, ETL workflows, and systems serving 10,000+ users. Now I'm pursuing my Master's in Computer Science in Germany, focused on machine learning: from RAG pipelines and LLMs to graph neural networks and explainable AI.
- 🔭 Currently working on the ARAG Data Pipeline - a modular Retrieval-Augmented Generation system with hybrid retrieval (semantic + keyword search) that grounds LLM outputs in verifiable sources to reduce hallucination and enable source attribution
- 🌱 Deepening my knowledge in NLP, knowledge graphs, and MLOps
- 💼 Open to working student roles and internships in ML/software engineering
- ✍️ I write and create educational tech content at TechiesTalk and on YouTube
- 📫 Reach me at faheemahmad.de@gmail.com
AI / ML
Backend & Frameworks
Data, Cloud & MLOps
| Project | Description | Tech |
|---|---|---|
| ARAG Data Pipeline 🚧 in progress | Modular RAG pipeline with hybrid retrieval (semantic + keyword), structured context augmentation, and grounded generation - improving factual consistency and source attribution over standalone LLMs | Python, RAG, Hybrid Retrieval, LLMs |
| Harvest Helper | AI-powered crop advisor: Gradient Boosting recommender (99.7% top-3 accuracy) + RAG pipeline with LLaMA 3.3 70B over FAISS, fused with real-time weather and Sentinel-2 satellite data | FastAPI, RAG, LLaMA, FAISS |
| Explainable GNN on Knowledge Graphs | R-GCN for node classification on the AIFB knowledge graph (91.7% accuracy) with interpretability analysis of model decisions | PyTorch Geometric, R-GCN, RDFLib |
| TapShip | e-Mandi platform connecting farmers directly to buyers - awarded a VTU innovation grant in agricultural technology | PHP, MySQL, JavaScript |
| Drowsiness Detection Device | Real-time driver drowsiness detection via eye-blink tracking with automated IoT SMS alerts | OpenCV, Computer Vision, IoT |
