Parth Parikh

Undergraduate Researcher

Aug 2025 – Present

STAR Lab (Georgia Tech)

• Streamlined end-to-end setup and debugging of TurtleBot4 hardware platforms, developing a ROS 2 control framework to publish commands and interface with onboard sensors for live multi-robot experiments
• Implementing single- and multi-agent reinforcement learning algorithms, including Vanilla Policy Gradient and Proximal Policy Optimization (PPO), to explore foundational techniques and performance trade-offs
• Developing reinforcement learning strategies for multi-robot teams, enabling agents to dynamically adjust control policies in real time for robust collaboration in changing and uncertain environments

Research Mentor & Teaching Assistant

Jun 2025 – Jul 2025

Carnegie Mellon University

• Advised a student research team developing a system to isolate individual voices from duet recordings, leveraging a Dual U-Net + BLSTM architecture and a custom dataset of 5,000+ duet samples, and guiding the project from ideation through data collection and analysis
• Mentored 70 high school students in Discrete Mathematics and theoretical computer science (state machines, introductory programming), as well as hands-on lab work with large language models, including dataset cleaning and fine-tuning TinyLlama using Hugging Face and Jupyter Notebooks
• Organized community-building and social activities to foster collaboration beyond the classroom

Data Science Intern

May 2025 – Jun 2025

Revvo

• Drove RAG chatbot development project enabling fleet managers and the customer success team to consume fleet-wide tire health insights through natural language, while maintaining development under Git-based version control
• Developed and validated 4 Java APIs and applied prompt engineering to optimize LLM response accuracy and JSON data formatting, securely linking internal Firebase and SQL tire health data with an external LLM
• Modified dice similarity-based matching algorithm to improve data input handling for domain specific terms (e.g. tire position nomenclature)

Software Intern

Sep 2022 – Jan 2023

Beaumont Technologies

• Resolved 3 data-visualization bugs in a polymer flow analysis tool for injection-molding simulations (C#), enhancing data accuracy and user clarity
• Implemented interactive graphing enhancements, including a hover-to-bold feature and a probing tool for targeted point plotting, improving client presentation clarity and satisfaction
• Collaborated with a plastics engineer and a part-time remote software mentor to identify pain points and design solutions aligned with end-user needs