Making Meta AI Agentic, enabling multi-step agent-driven AI responses on more product surfaces, with richer responses, and for more use cases.

Tech Lead on Google Maps cycling routing to provide worldwide users with a better and safer transportation experience. We are a team based in Tokyo but owns Google Maps Cycling completely collaborating with infrastructure platform teams and data quality teams all around.

• Determine technical roadmaps for year-long projects that involve cross-team collaboration, line up dependencies and set milestones;

• Lead design for data-serving infrastructure, delegate project components to team based on engineering bandwidth and team priorities, ensure consistent progress on deliverables;

• Drive implementation in routing backend, ship user-visible changes into production to users worldwide

Delivered Asia-specific features in driving navigation for Embedded Google Maps to Japanese automotive OEMs

• Determine guidance features with largest improvement headroom by conducting data analysis and case studies

• Align requirements from product management with engineering, to ensure external OEM's feature needs are satisfied and internal routing infrastructure can effective support

• Design and build voice guidance at complex intersections with irregular turns, highway entries / exits and tunnels (3D navigation), fine tune Japanese speech template for localization

• Collaborate with core experiment teams and platform engineers and shipped feature into production.

• Key architect of Ads Data Hub, a near-real-time data processing pipeline offering privacy-compliant data to Google ads customers in the new advertising age with no pixel tracking. Spent a whole year commuting between NYC and LAX to lead the engineering progress. Publicity: https://www.businessinsider.com/google-is-ramping-up-its-pitch-for-ads-data-hub-2019-11

• Launched a 10-quarter project in 3 quarters, providing a 6-hr streaming latency (compared to previously 30-hr) for ad metric data delivery to external advertisers and publishers via BigQuery.

Core architect and TL for Google's Unique Reach Audience Pipelines providing advertisers and publishers with low-latency ad reports of aggregated unique reach metrics sliceable in multiple dimensions.