M.S. Scientific Computing, Universität Heidelberg — NLP / Deep Learning / Generative AI
I build ML systems end-to-end, with a particular focus on the mathematical foundations and engineering rigor behind modern AI. My background in mathematics shapes how I approach problems: I care about understanding why something works, not just whether it does.
Portfolio: https://ameynarwadkar.github.io/
LinkedIn: https://www.linkedin.com/in/amey-narwadkar-474332231/
GitHub: https://github.com/ameynarwadkar
RAG Systems — full pipelines covering document ingestion, indexing, retrieval, and answer synthesis, with a strong focus on faithfulness, hallucination reduction, and retrieval quality (chunking, embeddings, reranking, grounded generation).
Generative AI — diffusion models, text-to-image generation, and practical LLM deployments.
Modern LLM Workflows — tool use, agentic systems, structured outputs, and evaluation frameworks.
Representation Learning — contrastive and self-supervised learning, embedding geometry, intrinsic dimensionality.
Generative Models — diffusion, multimodal learning, conditioning and control mechanisms.
Robustness and Evaluation — systematic evaluation pipelines, failure mode analysis, adversarial prompts, and calibration.
| Project | Area | Description |
|---|---|---|
| Tennis Analysis System | Computer Vision | Player detection, ball tracking, court keypoint estimation, speed metrics, and mini-court visualization |
| Text-to-Image Generation | Generative AI | Text-conditioned image generation built on a Stable Diffusion implementation |
| Food Ordering Chatbot | NLP | Conversational ordering flow with end-to-end user interaction |
| Sentiment Trading Bot | NLP / Finance | News sentiment-driven trading strategy with execution and backtesting pipeline |
| ML Algorithms from Scratch | ML Foundations | Core ML algorithms implemented from first principles using math and NumPy |
Languages: Python, JavaScript, R
Deep Learning / GenAI: PyTorch, TensorFlow, OpenCV, Hugging Face
RAG / LLM Tooling: LangChain, LlamaIndex, Pydantic, FAISS, Ollama
Data / Analytics: pandas, MySQL, MongoDB
Infrastructure: Docker, Kubernetes, FastAPI, CUDA, Google Cloud, Linux
I'm always open to conversations about RAG, GenAI, and NLP systems — whether that's collaboration, research discussion, or just trading ideas.
- LinkedIn: https://www.linkedin.com/in/amey-narwadkar-474332231/
- Portfolio: https://ameynarwadkar.github.io/