Shania Mitra
Biography
Shania Mitra is a Machine Learning Engineer at Volkswagen in Mountain View, where she develops LLM-based ADAS models and edge AI systems for real-time crash detection and decision-making. She is currently pursuing an M.S. in Computational Science and Engineering at MIT, working as a Graduate Research Assistant at CSAIL under Prof. Manolis Kellis on multimodal machine learning for protein function prediction, achieving significant performance gains over state-of-the-art methods.
An IIT Madras gold medalist with an M.Tech in Data Science and B.Tech in Chemical Engineering, Shania has also interned at Goldman Sachs and Bewgle Inc. Her research and projects span large language models, fairness in AI, healthcare analytics, and graph-based deep learning, with several publications at international conferences.
Research Interest
Shania Mitra?s research interests at the intersection of machine learning, large language models (LLMs), and multimodal AI. She focuses on developing LLM-based systems for real-time perception, decision-making, and edge AI applications. Her work integrates graph neural networks, transformers, and contrastive learning to advance protein function prediction and other computational biology tasks. She is also deeply interested in AI fairness and explainability, designing techniques to enforce fairness constraints in black-box models while preserving predictive accuracy. Additionally, her research extends to healthcare analytics, drug discovery, and efficient AI, where she develops low-latency, resource-optimized inference pipelines for scalable and robust applications
Abstract
