Building and maintaining highly reliable data ingestion system for complex financial datasets, developing libraries and middleware for data serving and LLM integration

• Overhauling ML model deployment pipeline, slashing launch times by 85%. Created a robust distributed key-value database handling millions of daily requests without a hitch.

• Pioneered AutoML for CV models, streamlining processes and cutting development time by 25%.

• Led the DevOps for 400+ ML models, ensuring zero down time during data center failure.

• Led the backend development of our CV labeling system. Boosted operational efficiency by 70% by introducing AI pre-labeling.

Built and optimized an in-house ads analytics platform from scratch, enabling dynamic data visualization and improving decision-making processes in collaboration with the ads team.