• Won Execution Award in AI Org(an honor to 1-2 individuals of the whole org of ~1000 people), and was a finalist a second time.

• Rich hands-on experience on LLM SFT & RL, RAGs etc.

• Pioneered the company's first end-to-end embedding-based retrieval, leading the efforts from model architecture design to system integration and deployment.

• Built highly efficient platforms in C++ to simulate large scale circuits and systems, achieving 10-100X runtime speedup and 10X less memory usage compared to commercial tools that rely on numerical methods, by building an automatic ordinary differential equation solver.

• Best paper nomination at a top-tier conference. Patent granted.