Sarthak Mahapatra

Research Engineer @Boston University Metropolitan College

Boston, MA, US

Developed mobile apps using Android Studio to test the Conjugate Gradient Descent (CGD) optimizer, demonstrating 2-4 fewer epochs for convergence and 10-20% reduced battery usage compared to the SGD optimizer.2. Implemented and fine-tuned 6-7 deep learning models across voice conversion, text classification, LLM fine-tuning, and image classification tasks using PyTorch and TensorFlow, applying CGD optimization across multiple AI domains.3. Conducted comparative analysis of CGD against PyTorch\'s built-in optimizers across large model deployments, showcasing CGD\'s superior efficiency for resource-constrained environments.

Boston University

Master of Science - MS, Applied Data Analytics

Delhi Technological University (Formerly DCE)

Bachelor of Technology - BTech, Electrical and Electronics Engineering

Delhi Public School Vasant Kunj

High School Diploma