Michael Brundage Email & Phone Number

Menlo Park, Ca, Us

Technical Lead and specialized algorithms expert for Machine Learning Understanding, AI Monitoring & Observability in AI Infra. I'm responsible for solving model stability, debugging, and explainability for Meta — especially fast detection, diagnosis, and root-causing of loss divergence and other training stability issues in our experimental and production ML models.

Menlo Park, Ca, Us

I was a Research Data Scientist (Quantitative Engineer) in Meta's Infrastructure Strategy group, a job role that no longer exists. I particularly enjoyed mentoring and growing the next generation of data scientists and managers.Projects included:• executive level metrics for AI Infrastructure. I implemented the first-ever structured metrics reviews at AI Infra VP level.• metrics and understanding for Ads Prediction Robustness. I proposed and created a "Revenue Loss Prevented" metric for ML snapshot validation which is now the APR top-line metric and resulted in additional RLP metrics in the years since.• the open source time series library, Kats. I was (and last I checked, still am) the top contributor after its public release.• misc projects for AI developer productivity, monitoring & observability; Ads Infrastructure; and Reliability Engineering.

Mountain View, Ca, Us

Core Developer Intelligence. Applied machine learning, analytics, and data science to understand and improve engineer productivity across all of Google.• Created predictive models for developer infrastructure• Created Developer Intelligence metrics and dashboards. Our work was mentioned by Ruth Porat in a NYTimes Conference and reported by NBC.• Internal owner of widely-used Python data libraries (seaborn, statsmodels, ...)• Presented invited keynote at internal anomaly detection conference* Helped YouTube team adjust their forecasts for the pandemic impact, and helped many other teams across Google (23 peer bonuses, 7 spot bonuses)• Volunteered at 2019 Women in Statistics & Data Science event in Bellevue, WA

Mountain View, Ca, Us

• First data scientist hired into Google's Supply Chain organization in Seattle/Kirkland.• Created automated ML forecast models for all of Google's Consumer Hardware products• Created the first weekly business review metrics decks for many teams• Taught metrics design and data science to hundreds of other Googlers• Created fancy data visualizations presented by our SVP on stage• Presented invited talk at JSM 2018, “What Statisticians Need to Know to Work in Tech”

Redmond, Washington, Us

• Shipped Windows 10• Led and advised various data science projects in the Windows and Devices Group (WDG)• Created and evaluated statistical methods to compare Windows quality and performance across builds• Architect for the WDG Experimentation Platform• Reviewer, Microsoft Journal of Applied Research• Session Chair of the Experimentation Track at the Machine Learning, Analytics, and Data Science (MLADS) Conference (Spring, 2016)

Various independent projects

Seattle, Wa, Us

• First engineer hired into Amazon Go.• Designed and implemented the initial technical architecture.• Created demonstrations for the CEO.• Represented Amazon in a visit to the MIT Media Lab.• Hired more than 30 Senior/Principal software engineers and research scientists.• Invented new approaches to hard computer vision challenges (more than 60 patent ideas).

Seattle, Wa, Us

I was responsible for the end-to-end worldwide consumer Search Experience, from the search box you type into to the results returned. This included both Amazon's web pages that display search results and the internal search service that powers the website and many other services including AWS.I wrote production code, invented technology, mentored other Principal Engineers and senior developers, drove cross-team issues, defined the technical direction for multiple teams, tackled company-wide technical challenges, and generally did whatever needed to grow revenues, reduce costs, and improve the customer experience.During this time, my team was directly responsible for more than $1B in incremental revenue, increased the fraction of Amazon Search pages that display the first screen of content in under 1 second from almost none to more than three-quarters, and decreased the failure rate by more than 1000x.I became a company-wide expert on website latency and front-end engineering, and frequently advised internal and external organizations on those topics. I invented click prediction, launched predicted keyword searches a month before Google Instant, and introduced AJAX to Amazon search pages.I was a coach for the Principals of Amazon talk series, in which I coached other Principal and senior Engineers in the weeks before their presentations to the entire Amazon engineer community. I also co-authored Amazon's Principal Engineer Tenets https://www.amazon.jobs/en/landing_pages/pe-community-tenetsI carried out the responsibilities of Data Scientist for the Amazon Search team for a year, which led to my helping create the Data Scientist career ladder at Amazon, now known as "Applied Scientist". I hired the first two Data Scientists at Amazon.

Sunnyvale, Ca, Us

I was the Distinguished Engineer on the Technical Product and Strategy team reporting to the CEO. We advised executive staff on technical strategy and overall direction for Yahoo!, opened and staffed a new Yahoo! Bellevue (Seattle) location, defined Yahoo!'s Web Services standards, and created the Yahoo! Open Strategy. In this role, I wrote software; evaluated and defined technologies, products, and platforms; assessed potential acquisitions; explored third-party partnerships; and connected the dots across the company. I invented the Yahoo Query Language (YQL), developed iPhone and Facebook apps (before those platforms launched), met with startup founders, and championed the formation of Yahoo's Cloud Computing and Data Infrastructure (CCDI) Division.

Redmond, Washington, Us

I was a senior software developer on the Xbox 360 launch team. Prior to launch, my team and I implemented the Xbox emulator for backwards compatibility, hailed as one of Microsoft's greatest technical achievements. I personally implemented emulation for the file system, networking and Xbox Live and created the backwards compatibility consumer feedback alias and FAQ.After launch, I became responsible for the Xbox 360 kernel/operating system and led the Xbox Performance (virtual) team. I worked on the multimedia client software for Xbox Live Video Marketplace and implemented MP4 and Zune support. I also implemented client-side instrumentation to gather data and report on console failures experienced by consumers.

Redmond, Washington, Us

I was a key contributor to Microsoft's XML query efforts, including adding XML to SQL Server (Microsoft's flagship database) and prototyping, designing, and implementing XQuery. An XML query algebra I invented became the basis for Office 2007's XML application framework and Microsoft's XSLT engine. I also represented the XML team inside and outside Microsoft. I wrote a book with Addison-Wesley, "XQuery: The XML Query Language," and co-authored a book with Wrox Press, "Professional XML Databases."

Pasadena, Ca, Us

I managed the software team for NASA's optical Interferometry Science Center, working on projects such as the Keck Interferometer and the Space Interferometer Mission. I was a programmer on AstroVR, a pioneering and influential pre-Web Internet application. I implemented an astrophysics simulator that was the world's fastest backwards ray-tracer and used to hunt for extrasolar planets.

Bs, Mathematics

Ms, Mathematics

Education record