Research Assistant
University of Florida
- Conducting a literature review on natural language to logic translation methods, synthesizing findings into a technical report shared with Dr. Sumit Jha's group.
- Benchmarking outputs from a 28K-pair NL2TL dataset utilizing LLM-based approaches, producing progress metrics that inform model refinement.
- Developing a baseline NL2TL translation system with LLMs, expected to increase the accuracy of temporal logic generation in preliminary testing by 15–20%.
Research Intern
Augusta University
- Collaborating with Dr. Yuyan Bao and PhD student Shichen Gao to integrate Knowledge Graph and LLM techniques while applying discrete mathematics principles, enhancing system accuracy, and preparing it to handle 1,000+ complex user-generated prompts.
- Assisting in multiple instance learning experiments, contributing to early results that showed up to a 12% improvement in recommendation precision.
AI Engineering Intern
SmartStickies
- Researched three transformer architectures and LSTM variants using 20K+ open-source interaction records, building predictive models that supported adaptive interface testing across multiple user scenarios.
- Designed and tested a Workflow Generator Tool that dynamically adjusted screen layouts, tested across three businesses, and integrated into production trials serving 100+ end users.
Python Programming Intern
Pinnacle Full-Stack
- Built three customizable Python applications through Tkinter, including a password generator, a weather forecast tool using the OpenWeatherMap API, and a translator app leveraging the GoogleTrans library.
- Streamlined API data handling and virtual environment setup, cutting average API response latency by 200ms and improving reliability during local testing in VS Code.