- As the first ML hire, I was responsible for ideating, designing, and creating the product architecture and its key features. Transformed a product with limited applications by adding machine learning capabilities, increasing its scope tenfold, and also developed several new features.- Developed and maintained the internal deep learning toolkit that uses PyTorch, OpenAI Gym and EnergyPlus to streamline workflow and research.- Built a Retrieval-Augmented Generation (RAG) LLM solution that processes and answers client queries on real-time data from an energy monitoring device using LangChain.- Achieved up to 30% reduction in energy consumption of HVAC systems using deep reinforcement learning.- Developed a PoC using deep reinforcement learning that saves up to 25% on electricity bills resulting in collaboration with a major industry player based out of Japan.- Optimized and deployed quantized neural networks onto microcontrollers for real-time industrialapplications.

• Utilised transformers and analysed travel data, aiding a city simulation research project and informing optimal urban mobility strategies.• Optimised travel routes for socio-economic impact with a combination of heuristics and deep learning with 92% accuracy.

Co-Founded VBIT's first Programming and Development Club. Conducted Competitive Programming contests and peer learning programs to inculcate coding culture in the college. Taught Data Structures and Algorithms in these peer learning programs.

Part of the Designing team of stuMagz's VBIT Chapter.