Experience

    • Engineered dbt pipelines in Snowflake to ingest and model 1M+ daily raw Snowplow events into Metabase
    • Constructed a unified semantic layer on Snowplow models, optimizing architectures to cut query runtime by 45%
    • Built a Claude-powered data agent for semantic views, enabling PMs to generate verified charts in under 60 secs
    dbtSnowflakeSnowplowMetabasePython
    • Developed a full-featured Microsoft Word Add-in using TypeScript and React, integrating a RAG pipeline to generate regulatory documentation drafts, increasing first-draft generation by 60%
    • Built source traceability, quality reports, and a RAG chatbot for regulatory documents, reducing review time by 35%
    • Automated cloud infrastructure with Azure Bicep, ensuring automatic, scalable deployment on Azure App Service
    TypeScriptReactRAGAzureBicep
    • Led the design of a personalized game recommendation system using Node.js, Express, and Steam Web APIs
    • Improved response time with Axios by optimizing API calls with caching mechanisms and parallel processing
    Node.jsExpressAxios

Involvement

    • Automated simulation test data generation and engineered pipelines to process and sync 10+ hours of teleop recordings
    • Fine-tuned pi0.5 and SmolVLA foundation models using flow matching to optimize real-world policy accuracy
    PythonPyTorchLeRobotIsaacSim
    • Developed cross-platform patient data management app with Electron and PostgreSQL, reducing data entry by 60%
    • Enabling Burmese doctors to track 200+ patients daily across 16 villages on the Irrawaddy river with offline software
    • Implemented offline storage and Excel export, reducing manual errors by 25% for outbreak monitoring & prevention
    ElectronPostgreSQLNode.js

Education

  • University of WaterlooBachelor of Mathematics, Applied MathematicsScientific Computing & Scientific Machine Learning
    Sep 2024 – PresentWaterloo, ON