Apr 2024 – Jun 2024

• Collaborated with traders and quantitative researchers to enhance trading strategies by adding new features and improving the process of analyzing daily trading results.
• Built a maker fill probability prediction model for trading strategies, leveraging changes in order books.
• Designed and implemented a trading log visualization tool, integrated into the post-trade pipeline for daily use.

Jun 2023 – Aug 2023

• Collaborated with the PTE Team to develop solutions for detecting crashes through user behavior simulation.
• Developed a bug report summarization tool using a suffix array algorithm to detect recurring log patterns, leveraging large language models for advanced pattern analysis and contextual inference.
• Engineered a user-centric Critical User Journey (CUJ) generation pipeline for crash detection, incorporating bug report summarization tools and optimizing paths with dynamic programming on directed graphs.

Jun 2022 – Jul 2022

• Migrated the CI/CD pipeline of the server management system from Drone CI to GitHub Actions for integration.
• Streamlined the process of automating server setup for infra team by Ansible, an open-source software tool.

Jun 2024 – present

• Advisor: Prof. Kexin Pei, UChicago Department of Computer Science
• Developed and experimented with a pipeline for generating stress tests for program performance evaluation,utilizing large language models to create test case generators based on given information.
• Conducted comparative studies on the effectiveness of different test case generators, analyzing their impact on runtime discrepancies across problems in different programming languages.

Sep 2023 – present

• Advisor: Prof. Pu-Jen Cheng, NTU Department of Computer Science and Information Engineering
• Designed multiple debugging pipelines and conducted fine-tuning experiments on open-source code-based language models, aiming to improve performance in competition-level program debugging.
• Built a dataset of 10K+ incorrect/correct solution pairs from real-world data for fine-tuning and evaluating models.

Feb 2019 – May 2020

• Advisor: Dr. Wei-Chao Chen, Chief Digital Officer & Senior VP at Inventec
• Implemented the SOTA encoder-decoder based model and trained it with a mandarin chinese voice dataset.

Feb 2024 – May 2024

• Led the organization of an inter-university programming contest, an ICPC-style competition involving around 30 teams, 80 contestants from the top three universities in Taiwan: NTU, NTHU, and NYCU.

Oct 2022 – Feb 2023

• Led a team of 50+ people, organizing a winter camp that trained 72 high school/college students with coding background for top-tier programming contests like the International Olympiad of Informatics (IOI).
• Raised USD$16000 from multiple firms such as JaneStreet, Systex, Mixerbox, and Sudo Research.

Mar 2021 – Aug 2022

• Organized an online programming education program by integrating the expertise of CS students to create a comprehensive remote learning experience, allowing students to continue their education during COVID-19.
• Led a team of 70+ people with the coordinator, managing the program for around 280 participants across Taiwan.