Conducting research under Prof. James Davis to create a framework for using LLMs to improve the energy efficiency of software running in data centers.Contributed to multiple parts of the framework:• Improved Chain of Thought (CoT) prompting, using self consistency and in context learning.• Verification of the optimized C++ algorithms through regression testing.• Evaluation of the framework using a standard software energy use benchmark• A feedback loop to collect energy data, create natural language feedback, and iteratively refine code for improved energy efficiencyCo-author for "Large Language Models for Energy Efficient Code: Emerging Results and Future Directions" (under review at the 2025 International Conference on Software Engineering)

Leading a team of developers to build an AI assistant which manages your inbox. Handles rule-based labeling, routine responses, forwarding and more.

Working on an integrated platform offering portfolio generation, simulated trading, and a chatbot for financial insights, using various data analysis methods.Using open-source transformer models (Financial Bert, Bart), APIs (OpenAI, yfinance) to conduct sentiment analysis, summarization tasks on web-scraped news articles.Worked with open source embedding models, vector databases like Chroma and LLMs to create an RAG-based bot, based on transcribed earning calls and quarterly reports.Won the JMEC Start-Up Incubator competition and was awarded $6000 to further the project.

• Developed an image processing solution in C++ to perform dust detection in low light. Optimized the overall image processing pipeline to bring latency down by 70% and CPU utilization down by 20%. Used OpenCV, MATLAB and Python for labeling data, experimenting with image features and visualizing results. • Worked in an agile framework within the See and Spray Select team to write unit tests for services like image remapping and dust detection. Learned to use Boost C++ libraries and CMake. Also gained exposure to CI/CD pipelines with the Jenkins automation platform.

Working on the Odometry sub-team to localize and implement simultaneous localization and mapping algorithms (SLAM) on an autonomous racing kart.Using ROS2, Python to implement odometry packages and SLAM algorithms based on 3D LiDAR data.Selected to present my work as a Research Talk at the Fall 2023 Undergraduate Research Expo.

Responsible for leading a weekly lab section with 30 students to help students understand fundamental concepts in C programming.Responsible for grading homework and weekly lab assignments, providing individualized feedback to students.

Created and led an organization with 16+ volunteers to develop a water conservation app and sustainability-based website. We conducted awareness efforts with corporate stakeholders and our public high school. In creating the app, I received technical mentorship through Xylem's Incubator Program and Tata Consultancy Services. The idea was recognized by Charlotte's NPR news channel (WFAE) and the National Diversity Council.

I led a team of 27 students in the business, build, and programming processes of creating highly competitive robots for the FIRST Robotics Competition. Responsible for guiding team meetings, outreach and recruitment efforts, and build/design for the team.