Nov 2020 – Nov 2025 · 5 yrs · Los Angeles, CA · Hybrid

• Led a team of engineers in the complete redesign and rewrite of Google's Ads Data Hub (ADH) core privacy engine, migrating it to a microservices architecture. This initiative reduced query privacy analysis time by up to 68% and virtually eliminated on-call alerts.
• Developed and implemented critical privacy protection mechanisms for the Personally Identifiable Information (PII) matching feature, which enhanced privacy checks and reduced unnecessary query rejections by a few percents. This required me to proactively upskill in C++ and work with a poorly documented service to write the majority of the project's codebase.
• Designed, implemented, and launched the Consent Administration service, a critical privacy feature, driving its end-to-end development and collaborating with multiple teams to resolve blockers and ensure a smooth, compliant rollout. This included coordinating with the Kids Infra team for notification replays and identifying and remediating a critical ADH consent collection bug with the UI team, ensuring data integrity and a successful feature rollout.
• Collaborated closely with enterprise customers to troubleshoot technical issues, support privacy-related query reviews, enable access to new event types on the Ads Data Hub platform, and guide integration of first-party or client-owned campaigns—contributing to sustained customer satisfaction and contract renewals.

May 2019 – Oct 2020 · 1 yr 5 mos · Vancouver, Canada Area

• Architected and developed a suite of data-driven services to provide customers with key insights into their virtualization environments. This included building services for on-demand PDF reports and a web UI for data visualization, as well as implementing a batch service to generate configuration recommendations for optimization.
• Implemented and maintained data collector components running on customer-managed infrastructure to gather VM health and performance metrics, transmitting telemetry to centralized cloud services for large-scale monitoring and analysis.
• Wrote a comprehensive internal guide doc about building and architecting microservices in Pure Storage using the Pure1 proprietary tool. Later this doc was used by several teams to develop their microservices and became one of the documents recommended by the SRE team.

Nov 2018 – Apr 2019 · 5 mos · Vancouver, Canada Area

• - Contributed to the multi-year development of a foundational subscription service for a new Salesforce Analytics Cloud initiative, serving a global customer base. Simultaneously maintained its legacy monolithic version to ensure a seamless business transition. The project delivered a scalable solution and a web UI for scheduled email reports.

Nov 2015 – Nov 2018 · 3 yrs · Vancouver, Canada Area

Dec 2014 – Nov 2015 · 11 mos · Vancouver, Canada Area

• - Built a backend service and reusable data visualization components for an analytics product to present data to clients. (JavaScript, ReactJS, D3)

Oct 2011 – Jul 2013 · 1 yr 9 mos · Montreal, Canada Area

• Master’s thesis (cited 74 times as of March 2025): Developed and evaluated supervised learning models to predict hospital readmissions within a specified timeframe, utilizing a large administrative health dataset. Incorporated medical taxonomies (drugs, procedures, diagnoses) as features, demonstrating improved prediction performance and reduced model complexity. Analysis revealed inherent limitations in readmission predictability, challenging the use of raw readmission rates as a sole indicator of care quality.
• Used a "mixed statistical model" to investigate the association between individual and social attributes of students of 20 high schools in the context of delinquent behaviors. Results showed that:
• 1) An average female student is less delinquent than an average male student.
• 2) The correlation between a student's delinquency & his/her friends' average delinquency becomes stronger as the time passes and it is even larger for more social students.