At HCL Technologies, I analyzed medical device engineering and product-performance data to support quality decisions. I also applied root cause analysis to product failures and worked to ISO 13485 and ISO 14971 standards.
I designed MedFusion-Leuk, an open-source multimodal leukemia decision support system combining blood smear images, lab parameters, clinical text, literature retrieval, and LLM explanations. Its image model reached 94.69% test accuracy, while its fusion classifier reached 100% accuracy on the stated evaluation.
I added SHAP and Grad-CAM explanations, RAG-grounded citations, and ICD-10/LOINC coding, then served the application through FastAPI and Streamlit. The project included 22 passing tests and CI/CD.
I’ve also built browser-based products including BounceSimulator, a deterministic physics and music simulator, and DocuVerse, a privacy-first document converter with AI tools. My work spans machine learning, clinical NLP, and end-to-end web products.

