Zeyu Edward Qi

About Zeyu Edward Qi

I’m a forth-year data science student at UC San Diego (Class of ’26), minoring in biology. I thrive on turning complex data into real-world solutions. My experience spans academic research, industrial software development, and teaching. Over the past two years, I’ve honed my skills in AI and machine learning—fine-tuning transformer models to boost inference speed, building end-to-end prediction pipelines, and deploying robust ML-driven applications. I complement this with strong exploratory data analysis chops, using Python, SQL, and visualization tools to unearth actionable insights from messy datasets.Currently, I am working as a student researcher at UCSD\'s Living Matter Lab. I have been working on several projects. In Project Mitotic Wave, I use TIF videos of sea star embryo cells to locate the position of nuclei through MIP projection, track cell division using TrakeMate and visualize by labeling cells with generations, and calculate mitotic wave by using Delaunay Triangulation algorithm. In Project Morphogenesis, I captured time-lapse videos of sea star embryo cell division and stabilized cells for analysis through correcting movement artifacts by using translation and rotation adjustments along the X and Y axes. Also, I utilize Python to import and manage scientific packages for image processing, by analyzing ciliary movement wavelengths from video data to generate preliminary insights into cilia-driven flow patterns.I am interested in the field of bioinformatics, biostatistics, and data science. I’m excited to connect with professionals in biostatistics and biotech who are advancing data-driven healthcare. Let\'s conncet!

Experience

1. Data Scientist

   Chartgen Ai

   Sep 2025 — Present · San Diego, CA, US

   Integrated Ada.im’s LLM into the DABstep benchmark by building a custom provider shim, extending the model registry, and automating evaluation across ~450 multi-step data analysis tasks- Developed post-processing pipelines to normalize natural-language outputs into strict JSONL schema (scalars/strings), reducing formatting errors and enabling leaderboard submission- Configured trace logging and concurrency workflows (smolagents, LiteLLM, Arize Phoenix) to improve reproducibility, debugging, and team adoption of benchmarking results.

Education

• UC San Diego

  Bachelor of Science - BS, Data Science