SHUNYU YU

Fine-tuned a CLAP audio-language model on 33,813 bird audio-text pairs across 502 species for natural-language retrieval, owning the full data and evaluation pipeline and reaching a 21x improvement over base CLAP with zero-shot generalization to unseen species.

Developing software-in-the-loop simulation frameworks and flight control systems for fixed-wing UAVs.

Proposed and built an internal QA platform with React, FastAPI, and PostgreSQL that digitizes handwritten forms with an OpenCV and dual-model PaddleOCR pipeline, cutting form processing time about 40 percent, with CI/CD on GitHub Actions, Docker, and AWS.

Developed firmware and electrical systems for mobility solutions, focusing on motor control and telemetry

Assembled electrical cabinets and systems, ensuring compliance with safety and engineering standards

Curated and produced engaging editorial content highlighting developments in space science and exploration

Provided technical support for IT systems in a workplace environment