At Oasis Infobyte, I developed data-driven solutions using Python and statistical modeling. I received a Letter of Recommendation and was recognized as a “Star Performer” by the AICTE–Oasis Infobyte Data Science Internship Program.
I built a financial fraud detection system that modeled 50+ accounts as a transaction network. Degree centrality helped flag the five highest-risk accounts, and automated CSV reports turned graph metrics into a suspect list.
For rag-qa-system, I built a document-grounded RAG API using FAISS and Gemini 2.5 Flash, achieving sub-50ms retrieval latency. I also designed asynchronous document ingestion and tested strict-grounding prompts against direct, missing-information, and partial-answer queries.
I built NeuralLink-AI, a real-time sign-language recognition system using stacked LSTMs and MediaPipe hand landmarks, and BioSync-AI, a contactless heart-rate monitor using facial landmarks and signal processing. My work spans machine learning, deep learning, computer vision, and GenAI systems.

