Developed energy price forecasting models across multiple ISOs using ML techniques for structured time series data, experimenting with features and model selection, driving significant P&L uplift in trading, and enhanced EQ’s internal ML forecasting framework for the energy grid.

Partnered with research scientists, quants, and traders to productionize and launch new trading strategies, improving market coverage and execution reliability.

Built and maintained the trading operations platform to perform risk and compliance checks, ensure timely trade submissions to markets, configure alerts, and respond to incidents.

Worked on several large-scale, critical privacy software systems to ensure Meta’s compliance with various privacy regulations.

Designed & implemented several improvements to the Privacy Jurisdiction service (PJ), which classifies billions of users based on which privacy regulations apply to them, and is used by ~200 features across Meta to control how product and system logic changes for users/entities based on which privacy regulations apply to them.

Developed the Consent Data Platform (CDP), a scalable logging and continuous data monitoring system to monitor the health of ~100 consent flows that span across 4 apps and 14 surfaces.

Designed a continuous verification system to identify cases where consent flows were not reaching eligible users, or where user consent was not being respected, and helped with remediation.

Led the data monitoring efforts for the launch of Meta’s response to the POPIA, PDPA and PIPA regulations, which consisted of collecting user consent for usage of their data through consent flows and ensure eligible users could be brought to a compliant state.

Built features to help small businesses using FBLite (FB app for users with poor connection or low-end devices) grow their online business presence, such as making it easier to promote content for free without paying for ads by allowing business pages to join FB groups and market their products/services organically, and suggesting relevant groups in which they could post content.

Improved the performance of the FBLite business surfaces by reducing time for a business page to load.