Aditya Shelke

Studying computer science and mathematics at the University of Maryland, College Park. Minor in robotics and autonomous systems.

EXPERIENCES

Bloomberg (Summer 2025)

Designed and built a distributed system of controllers for infrastructure autoscaling with Golang, Postgres, Azure.

Bank of America (Summer 2024)

Built software to automate trade order management and modification with Python and NLU. Developed socket - based API orchestration software.

Metron (Summer 2023)

Developed a multithreaded Bayesian sampling system in Java improving SONAR tracking efficiency by 50%. Implemented modern MCMC and adaptive rejection techniques, cutting sampling errors by 75%.

PROJECTS

Future Frame Image Segmentation

Designed a vision-inertial learning pipeline for autonomous vehicles that fuses RGB sequences and IMU motion data using LSTMs, FiLM conditioning, and ConvLSTMs, to predict drivable area at a future timestep.

BitTorrent Client Implementation

Implemented a BitTorrent peer-to-peer client in Go, supporting TCP/UDP trackers, IPv6, tracker refresh, and concurrent downloading/uploading using goroutines and a producer - consumer model.

Facial Recognition with Triplet Loss

Implemented a facial recognition system using triplet loss in PyTorch, achieving high accuracy on benchmark datasets.

Neighborhood Graphs for Research

Developed an application to map significant relationships between published papers. Additionally created a paper ranking and relevance search algorithm.