•Developed an interactive neural network model editing platform using React, TypeScript, ReactFlow, and Dagre, enabling users to modify computational graph structure and layer parameters via drag-and-drop interactions.
•Integrated existing model inference cost estimation APIs by building Python/Flask end points and visualization components, allowing users to evaluate latency, FLOPs, and optimization impact after model edits.
•Implemented export functionality for TVM and JSON model formats, supporting downstream model compilation and deployment workflows.
•Collaborated with ML and software engineering teams, conducted testing, and maintained documentation to support scalable ML tooling development.
