PhD thesis 'Data Centric Workflows for Crowdsourcing Applications' awarded best thesis prize at BDA2021.
Research Experience
During PhD, member of DRUID and SUMO team, focused on developing algorithms for crowdsourcing applications, proposing formal models for complex workflows, and developing human-in-the-loop algorithms using probabilistic techniques and latent variable models. Before PhD, worked as a Researcher at TCS Research & Innovation Lab, Kolkata, India, working on various problems related to healthcare and sensor data (outlier detection, noise cleaning, disease classification).
Education
PhD from INRIA, IRISA research center, Rennes, France; Thesis: 'Data Centric Workflows for Crowdsourcing Applications', Co-advised by Dr. Loic Helouet and Dr. Zoltan Mikols. Master's from IIT Patna, Computer Science Department, June 2015; Master's Thesis: 'Push based User Selection in Crowdsensing', Supervised by Dr. Ashok Singh Sairam.
Background
Senior Chief Engineer at Samsung Research, Bengaluru, India. Loves to train large language and multimodal models with a focus on Generative AI in Reasoning Systems, Virtual Assistants, and recently agentic-based systems. Fascinated by the potential of these models to understand and reason. Aims to make advanced models more accessible by optimizing them for edge devices using techniques like distillation and pruning.
Miscellany
Open to new opportunities, projects, teaching, and collaborations.