Leading engineering team working on Location for Search and for other products at Google.~35-person cross-functional team, focused on machine learning, search quality and full-stack development across Google's codebase. The team runs several millisecond-latency and million-qps services.Strong focus on clear and crisp internal communication, cross-team influence and stakeholder management, in an area under heavy regulatory scrutiny (GDPR, Digital Markets Act).

Led Maps Timeline engineering team at Google. Transparency and value around your Location History. ~20-person full-stack and cross-functional team.Defined 3-year strategy and vision for the team working with very sensitive data from large user base, in a complex privacy and regulatory environment. In 3 years, doubled the monthly usage of Timeline, an already mature product. Launched overhaul of technical infrastructure, as well as a few sustainability features.Coached individual contributors and managers on Google’s performance evaluation system, reaching 2000+ engineers and managers.

Leading Geocoding team at Google. Mapping search queries (e.g. addresses, cities, countries) to their location on Earth. ~30-person team.Built up full-stack team: search quality, metrics, frontend, on-device/offline geocoding, developer APIs. Hired key leaders for new investment areas on the teamInvested in integrating with KnowledgeGraph-based search backend. Successful 2-year project with strong quality gains and huge system simplification.Established collaborations with maps data teams and with the rest of local search quality at Google. Participated in defining new, scalable metric to capture state of maps quality in next-billion-user countries, where traditional methods (paid raters or golden query sets) cannot work.Areas of interest include: search quality, information retrieval, data quality, metrics, frontend development, address processing, on-device systems (Google Maps running on your phone), developer APIs.

Tech Lead/Manager of Geocoding Search Quality: re-started and grew team answering queries for addresses (e.g. [1600 Pennsylvania Ave]) on google.com and maps.google.com.Improved 2% of maps queries every year, from completely broken to perfect (reduced search quality headroom by 10-20% per year).Accelerated move to Machine Learning for ranking/scoring of answers, resulting in huge quality improvements.Established metric for side-by-side comparisons and continuous tracking of maps search quality; used by 50+ software engineers as their main yearly optimization target; used in Google-wide web search launch process.Grew tech leads and managers on the sub-teams.

Tech lead and manager of Geocoding Data Quality: parsing addresses from external partners (e.g. business listings) and on web pages.Scaled address understanding of partner feeds to 100M+ listings a day and reduced processing time of largest feeds by 60%. Improved address annotation recall on web pages by 5% and decreased CPU usage of the address annotator by 40% (savings: ~1000cpus)

Individual contributor and Tech Lead on Location Extraction Service (100+ kqps, running early in google.com and maps.google.com search flow).Designed and implemented hot-swapping of search index for emergency fixes. Designed and implemented direct in-memory loading of binary and index, which reduced deployment of new versions of the software from 20h to 45min.

Developed communication protocols for fault tolerant systems in SDL (Specification and Description Language). Integrated subsystems written in SDL and Java.