About

Work Experience

• Head of GCP Quantitative Research & Generative AI Evaluation, Google

  Sep, 2020 - Present

  Directing an organization of 13 data scientists and quantitative researchers, defining enterprise evaluation frameworks, causal measurement science, and telemetry architecture for Google Cloud’s ~$30B+ ARR portfolio.

  • Invented the Specification-Driven Agent Evaluation (SAgE) framework, transitioning non-deterministic Multi-Agent System (MAS) evaluation from heuristic string matching to formal state machines with precondition, invariant, and postcondition verification.

  • Authored Google Cloud’s foundational Measure AI Playbook, awarded the Alphabet-wide 2026 GUXIES Honorable Mention in UX Research (selected from ~100 competing AI initiatives) for driving a 2.5x customer satisfaction lift in Gemini Cloud Assist (GCA).

  • Engineered the production skill evaluation harness, Turtle (.ttl) Ontology Visualizer, and automated Everest/EvalHub pipelines executing continuous evaluation runs across 5+ AutoCloud agent tiers to guarantee deterministic release gates.

  • Modernized Core User Journey (CUJ) tracking across an 8M-line legacy codebase, slashing critical workflow complexity by 66% (9 steps to 3) and deploying 4 AI analytical tools into production.

  • Established the Data Transition Command Center post-reorg, actively safeguarding and homing ~6,000 data assets and ~4,000 research artifacts across 30+ product teams with zero telemetry downtime.

  • Engineered the backend classification infrastructure for the LLM-Verbatim Classifier with automated PII redaction for 14,000+ users, unblocking Gemini HaTS-in-Chat (Happiness Tracking Surveys).

  • Modeled infrastructure adoption for Extended Custom Constraints (ECC), scaling to ~1M daily VM core-hours and boosting Cloud Load Balancing adoption by +8%.

• Senior Data Science Manager, Microsoft

  Jan, 2016 - Sep, 20204 years 8 months

  Led a cross-functional data science organization to shift product strategy for a $6B business portfolio from intuition-based to predictive, data-driven decision engines.

  • Architected Power BI’s long-term data strategy, moving from reactive reporting to predictive, ML-driven customer segmentation that forecasted user needs and marketplace behaviors.

  • Established 10+ quantitative analysis programs focused on customer health, retention elasticities, and usage behaviors across the Power Platform ecosystem.

  • Won the 2019 Microsoft Global Hackathon (Microsoft Garage) for developing an AI for Education prototype integrated into Microsoft Teams.

• User Experience Researcher & Data Scientist, Microsoft

  Sep, 2012 - Jan, 20163 years 4 months

  Mixed-method UX researcher on Windows, Phone, and Xbox, conducting field studies, usability benchmarking, telemetry log analysis, and large-scale experimentation.

  • Scaled a usability benchmarking engine (team of 10) executing 50+ weekly sessions, accelerating release cycles and establishing core experience quality standards.

  • Translated foundational behavioral models and theories into product designs for Windows 8, 8.1, and 10, impacting hundreds of millions of users globally.

  • Drove Windows Phone East Asian product strategy through international competitive research and mixed-methods data triangulation.

• Visiting Research Scientist, Palo Alto Research Center (PARC)

  Jun, 2010 - Jun, 20122 years

  Designed and implemented Sense Places Out of Text (SPOT), a geo-disambiguation algorithm and NLP annotation platform (US Patented).

  • Invented patented natural language processing and spatial entity extraction algorithms for unstructured text streams.

• Research Intern, IBM Research Center

  Jun, 2009 - Sep, 20093 months

  Developed a statistical predictive model to analyze relationship closeness across ~60K IBM employees, nominated for Best Paper at ACM CHI 2010.

  • Published novel social network closeness models at ACM CHI 2010 (Best Paper Nomination).

• Research Intern, IBM Research Center

  Jun, 2010 - Sep, 20103 months

  Conducted user studies (interviews, log telemetry, online surveys) for enterprise mobile applications, directly guiding product development.

Skills

Education

Awards

Volunteer Work

Publications