i'm Shu, an AI generalist passionate about open source and exploring the intersection of technology and design.
i'm currently working on ProteinGym - an open source benchmarking platform for evaluating machine learning models for proteins. this project is a combination of computational biology and machine learning applications in protein engineering, with proteingym-base to process data and models and proteingym-benchmark to benchmark selected data and models.
here are some projects i've worked on in the past:
- an interactive data visualization project exploring color theory and design principles.
- tech stack: SvelteKit, D3.js, TypeScript, Tailwind CSS.
- a distributed network traffic analysis system that captures, traces, and visualizes network topology patterns using graph-based visualization, inspired by "the last of us".
- tech stack: Apache Kafka, ksqlDB, Kubernetes, Helm, Neo4j, Scapy, Neovis.
- a data pipeline project demonstrating how to migrate from Excel-based workflows to a structured database system with interactive querying and visualization capabilities.
- tech stack: PostgreSQL, Apache Superset.
- a computer vision research project analyzing human poses and body contours in classical and modern artwork using pose estimation and body segmentation techniques.
- tech stack: OpenPose, DensePose, COCO Dataset, AdaIN Style Transfer.
