We're building a platform that empowers anyone to build powerful applications and workflows directly on top of their company's data. Imagine creating your own version of tools like Anaplan or Expensify, without needing to write code, all within Sigma.
Building Sigma's AI copilot for workbooks (AI Builder) — an agent that uses natural language to create dashboards and applications. I helped design the core agent harness and helped build the evaluation infrastructure now used across all of Sigma's AI features.
I also led the App Logic team, building Actions — a low-code solution that transforms static BI dashboards into interactive workflows with background execution and AI integrations.
Deployed ML models to predict potential hardware and software failures in real time for over 23,000 internet devices that power the Azure network, aiming to increase overall reliability
•
Incrementally trained XGBoost models on over 40GB of data daily by implementing batch training methods to prevent model drift and increase overall performance
•
Generated new features by leveraging feature engineering and scheduled SQL functions to increase model performance to over 90% accuracy
