Loading...
Total Website Visits
Hello, World! 👋
Welcome to my resume! I'm a passionate computer science graduate student with a strong background in artificial intelligence, robotics, and software engineering. I thrive on developing innovative solutions using cutting-edge technologies and collaborating in diverse teams.
With a perfect 4.0 GPA from Georgia Tech and a stellar academic record, I'm ready to take on new challenges and make a lasting impact in the tech world. Let's embark on an exciting journey together and create something truly remarkable!
Education
Georgia Institute of Technology
Atlanta, GA, USA
Master's in Computer Science, Specialization: Perception & Robotics
GPA: 4.0/4.0
Aug 2022 - May 2024
Manipal Institute of Technology
Manipal, KA, India
Bachelor's in Computer Science, Specialization: Artificial Intelligence
GPA: 8.97/10.0
Aug 2016 - May 2020
Experience
Graduate Student Researcher
CORE Robotics Lab
May 2023 – May 2024
- Investigating multi-agent task allocation and scheduling problem by formulating it as an NP-hard optimization problem
- Implemented Heterogeneous Graph Attention networks and Heterogeneous Graph Transformers for learning the agent and task embeddings for generating near-feasible schedules
- Formulated a Markov Decision Process (MDP) and performed reward engineering for training the scheduling policy using Proximal Policy Optimization
Graduate Teaching Assistant
Georgia Institute of Technology (Intro to Perception & Robotics)
Jan 2023 – May 2024
- Developed programming assignments and docker for them to deploy on the Autograder for evaluating concepts on exploration using Frontier Planning, motion planning using probabilistic sampling methods and dynamics controller using Proportional-Integrator-Derivative for 350+ students
Systems/Software Engineer
Hewlett Packard Enterprise
Jan 2020 – Jul 2022
- Enhanced Java Database Connector for Nonstop SQL/MX database by adding support for large object data type in ResultSet.Sql APIs, improving query performance and eliminating the need to distribute large objects across multiple rows
- Designed, developed, and automated tests for qualitative assurance of large-object (LOB) data-type support and concurrent database operations (DDL/DML) via Java Database Connector using JUnit and ApacheAnt framework
- Successfully led the migration of an internal developer's tool (Windows application) to the latest Visual C++ and SSL libraries, collaborating with and supporting four development teams throughout the migration process
- Developed a proof of concept for improving the performance of NonStop SQL/MX database execution engine by enabling high volume inter-process communication data transfer. This increased the inter-process packet from 32KB to 2 GB.
Projects
Real-time Balance Calculator Service using Event Sourcing
Java, Spring Boot
Mar - April 2024
- Developed a real-time balance calculator service with REST APIs for authorized fund additions and deductions.
- Maintained an event log to track all the events and enable server state restoration in case of failure.
Multi-Agent RL Benchmark and Learning Environment for Robotarium (MARBLER)
Python, Pytorch, Robotarium
Jan - Jun 2023
- Developed an open-source OpenAI Gym environment for benchmarking Multi-Agent Reinforcement Learning (MARL) algorithms.
- Supported sim-to-real transfer to turtlebots through Georgia Tech's Robotarium simulator.
- Performed extensive hyper-parameter search for benchmarking SOTA MARL algorithms implemented in the E-PyMARL platform on MARBLER.
Publications
MARBLER: An Open Platform for Standardized Evaluation of Multi-Agent RL
Best Paper Award MRS 2023
- Developed MARBLER (Multi-Agent RL Benchmark and Learning Environment for Robotarium), an open-source OpenAI Gym environment for benchmarking Multi-Agent Reinforcement Learning (MARL) algorithms
- MARBLER supports sim-to-real transfer to turtlebots through Georgia Tech's Robotarium simulator, enabling comprehensive and standardized evaluation of learning algorithms on realistic simulations and physical hardware.
- Performed extensive hyper-parameter search for benchmarking state-of-the-art MARL algorithms implemented in the E-PyMARL platform on MARBLER.
Analysis of Calibration Techniques for COTS Sensors Onboard a Nanosatellite
AeroConf 2019
- Researched and compared methods for characterizing and calibrating Commercial Off-The-Shelf (COTS) sensors onboard a 2U class nanosatellite, focusing on Anisotropic Magnetoresistance (AMR) Magnetometer and Micro-Electro-Mechanical Systems (MEMS) Gyroscope.
- Developed a mathematical model to account for internal errors and external factors affecting sensor values, and implemented an offboard neural network using backpropagation to define a nonlinear relationship between raw sensor values and actual values.
- Analyzed and compared the calibrated values obtained from both methods, presenting the results through graphs and critical analysis.
Design and implementation of a reaction wheel system for Cube-sats
AeroConf 2018
- Developed the dynamics and control of a reaction wheel system for three-axis control of a nanosatellite, including the dynamics equations, controller-reaction wheel interface, and motor control mechanism.
- Designed a controller-reaction wheel interface that includes speed control of the reaction wheel motor based on obtained torques.
- Implemented a momentum dumping strategy using three magnetorquers to unload momentum from saturated reaction wheels for continued attitude control.
- Tested the speed control system and momentum dumping strategy using a Software-In-the-Loop (SIL) system, presenting relevant results.
Technical Skills
Languages
- C/C++
- Java
- Python
- SQL
- HTML
- CSS
Libraries
- PyTorch
- scikit-learn
- Pandas
- Numpy
- cleanRL
- spinningUp
- networkx
- Deep Graph Library(DGL)
- CUDA
- MPI
Tools & Technologies
- wandb
- Docker
- ROS2
- Visual Studios
- JUnit
- PyTest
- MATLAB
- Apache Ant
- Git
- Docker
- Make