I'm a software engineer with interests in machine learning, math, finance, and philanthropy. I formerly worked on Microsoft's Bing search engine, improving the relevance of its results.
FlyHomes (2015 ‑ present)
- Built from scratch a machine-learning model to predict the value of homes in Seattle. The model's accuracy was within a few percentage points of Zillow's predictions.
- Created the company's data-processing pipeline, from downloading a dozen different data sources to inserting combined, cleaned entries into the website database, with data refreshed hourly.
Foundational Research Institute and Essays on Reducing Suffering (2013 ‑ 2015)
- I helped to create, organize, and run the Foundational Research Institute (FRI), a nonprofit organization. I write essays and scholarly articles on behalf of FRI. I fundraised the entire 2015 budget and hired our first full-time employee.
- I've written over 130 essays for my website, which gets 300 unique visitors per day and is widely cited as an important source of ideas for the burgeoning effective-altruism movement.
- My writings helped to jumpstart growing interest in reducing wild-animal suffering. As a result of my work, several academic papers have been published on the topic, Peter Singer has cited my essays, and a Facebook group that I created on this issue has reached 1200 members as of May 2015.
- I was one of the first in the effective-altruism movement to describe the argument for "earning to give."
- I built and maintained the websites of FRI and Essays on Reducing Suffering. I created the FRI podcast.
Microsoft, Bing, Core Ranking (2009 ‑ 2013)
Software Development Engineer II
- I designed and evaluated new data-mining techniques and algorithms for improving the relevance of search results. This ranged from basic statistical analysis and feature visualization in Excel to larger experiments using sophisticated machine-learning tools and software, as well as processing on terabyte-sized data sets.
- I built end-to-end three of Bing's production ranking algorithms by combining techniques from the rest of the team. Each one took several months of preparation, experimentation, debugging, and qualification. In the end, these were the rankers that served over 5 billion searches on Bing and Yahoo every month.
- I informally led small teams in shipping six major sets of ranking techniques in a variety of dimensions: Training data, popularity signals, feature engineering, query classification, and result freshness.
- I envisioned, planned, built, tested, and shipped a major breakthrough in Bing's ranking of one of its worst-performing freshness segments, resulting in huge relevance gains and a new platform for innovation in that space.
- I helped to teach dozens of teammates tricks of the trade for ranking, and I was a primary mentor for five colleagues, helping them to ramp up as successful engineers of their own.
Watson Wyatt Worldwide, Washington, D.C. office, Retirement Division (summer 2007)
Actuarial Analyst Intern
- I completed a study of the accuracy of a client's actuarial retirement assumptions and delivered a presentation of the results to the Retirement actuary division.
Washington, D.C. Office of Congressman Michael R. McNulty (summer 2006)
- I led tours of the US Capitol, wrote responses to constituent letters, and prepared briefs on committee hearings.
New York State Labor-Religion Coalition (summer 2005)
"Sweatshop-Free Schools" Intern
- I researched and compiled a list of sweatshop-free manufacturing companies for use by organizers throughout New York State.
Swarthmore College, Swarthmore, PA (2005 ‑ 2009)
- GPA: 3.97 out of 4.0.
- Major: Computer Science.
- Minor: Mathematics/Statistics.
- Graduated with High Honors.
- Academic Transcript (pdf)
- Relevant coursework: Probability, statistics, machine learning, machine vision and audition, natural language processing, information retrieval, data structures, algorithms, database systems, discrete mathematics, linear algebra, ordinary and partial differential equations, real analysis, mathematical finance, microeconomics, macroeconomics, accounting.
- Awards: Phi Beta Kappa, Sigma Xi: The Scientific Research Society, Federal Academic Competitiveness Grant for academic excellence, Merritt Hallowell '61 Endowed Scholarship for high academic achievement.
Guilderland Central High School, Guilderland, NY (2001 ‑ 2005)
- GPA: 4.0 out of 4.0. Valedictorian from a class of roughly 500 students.
- Awards: National Merit Scholarship, Robert C. Byrd Scholarship, National AP Scholar Award.
- GRE (2008): Quantitative: 800/800, Verbal: 730/800, Writing: 6/6
- SAT (2004): Math: 740/800, Verbal: 780/800
- ACT (2004): 34/36
- Passed actuarial exams (2006-07): P/1 (probability) and FM/2 (interest theory and derivatives).
- Computer: Python, C#, Visual Studio, Microsoft Excel, HTML, WordPress, Google AdWords. Have used Linux, Mac, and Windows. (my GitHub page)
- Languages: Studied German for seven years.
- Brian Tomasik, "A Minimum Description Length Approach to Multitask Feature Selection," 2009 (local pdf copy).
- Paramveer S. Dhillon, Brian Tomasik, Dean Foster, and Lyle Ungar. "Multi-Task Feature Selection using the Multiple Inclusion Criterion (MIC)." ECML PKDD '09. (Matlab code, GitHub version)
- Brian Tomasik, Douglas Turnbull, Joon Hee Kim, Margaret Ladlow, Malcolm Augat, Derek Tingle, and Richard Wicentowski. "Using Regression to Combine Data Sources for Semantic Music Discovery," ISMIR '09.
- Brian Tomasik, Phyo Thiha, and Douglas Turnbull. "Tagging Products using Image Classification," SIGIR '09.
- Brian Tomasik and Dougal Sutherland, "An Efficient Python Module for Lexical Distributional Similarity," 2008.
- Stephen S. Golub and Brian Tomasik, "Measures of International Transport Cost for OECD Countries," OECD Economics Department Working Papers, 2008. Prepared as part of a larger OECD study of "The Contribution of Economic Geography to GDP Per Capita" by Hervé Boulhol, Alain de Serres, and Margit Molnar.
Animal Charity Evaluators (2012 ‑ 2015)
- I helped advise founding, website creation, and initial hiring.
- I provided feedback and oversight as a formal Board member.
Various nonprofit organizations (2012 ‑ present)
- I helped set up Google Grants campaigns for four nonprofit organizations.
Donations (2009 ‑ present)
- I've donated ~$200,000 to charities I support.
Wikipedia contributor (2013 ‑ present)
- I've written 16 new articles and added to dozens more. See a list of my contributions.