# Vincent T. > Software Engineer at Google DeepMind Location: San Francisco, California, United States Profile: https://flows.cv/vincentt ## Work Experience ### Software Engineer @ Google DeepMind Jan 2021 – Present | Mountain View, California, United States Working on enabling programmers to effectively deploy machine learning in large-scale production systems, with a focus on reducing engineering effort and scope for errors. See technical details: https://arxiv.org/abs/2304.13033 ### Senior Software Engineer @ Cruise Automation Jan 2020 – Jan 2021 | San Francisco Bay Area - Productionized measurements of simulation fidelity, eliminating inconsistencies & reducing engineering effort by ~90%. - Increased company-wide velocity by eliminating ~1 million unused lines of code. - Saved ~2000 hours of developer time via a comprehensive dashboard offering real-time insights into recent test results. ### Software Engineer @ Cruise Automation Jan 2018 – Jan 2019 | San Francisco Bay Area - Led team of four to optimize simulation code, reducing runtime by 50% & operating costs by ~$5M/yr. - Developed a user-friendly framework for maintaining backward & forward compatibility between different versions of serialized data, reducing user error by ~90% and total LOC by ~80%. ### Research Assistant @ Robot Locomotion Group, MIT Jan 2017 – Jan 2018 | Cambridge, MA Developed a fast verifier (two to three orders of magnitude faster than the state of the art) to evaluate the robustness of neural networks to adversarial examples. The computational speedup enabled: 1) Verification of properties on convolutional and residual networks with over 100,000 ReLUs 2) Computation (for the first time) of the exact adversarial accuracy of a classifier subject to bounded perturbations Across all networks and datasets considered, we certified more samples than the state-of-the-art and found more adversarial examples than a strong first-order attack. Analysis of factors affecting verification time led to follow-on work creating a regularizer to further reduce verification times. ### Software Engineer Intern @ Palantir Technologies Jan 2017 – Jan 2017 | San Francisco Bay Area Reduced worst-case latency of a versioned datastore by 10× by optimizing costliest code paths, enabling efficient access to terabytes of data. ### Undergraduate Student Researcher @ Robot Locomotion Group, MIT Jan 2015 – Jan 2017 | Cambridge, MA Developed a fast and robust reactive controller for obstacle avoidance by: 1) Designing representative simulated environments 2) Collecting training data by sampling points in these environments, and computing at each point a) the simulated sensor readings and b) the control input selected by a controller with full knowledge of obstacle positions 3) Training a linear model to map sensor readings directly to control input, without estimating obstacle positions ### Software Engineer Intern @ Benchling Jan 2016 – Jan 2016 | San Francisco Bay Area - Designed and implemented real-time collaborative document editing with React, Flux and PostgreSQL for over 3,000 weekly active users. - Implemented several highly requested features for the lab notebook, including printing a selection of a document and enabling insertion of rows and columns in tables, in under two weeks each from start to release. ### Software Engineer Intern @ Jane Street Jan 2015 – Jan 2015 | New York, NY - Extended core OCaml library with a real-time stream processing package. Package enabled development of reusable components by simplifying error handling and control flow. - Designed and implemented a system for high-frequency writes and low-latency queries on time-series data, enabling better monitoring and reporting of data center operations. ### Research Assistant @ MIT Sloan School of Management Jan 2014 – Jan 2015 | Cambridge, MA - Studied how equilibrium returns for a firm are affected by firm entry dynamics and risk linked to foreign competition. - Validated and standardized data sets with ~100,000 datapoints; work formed basis for a published paper. ### Data Science Intern @ EnerNOC Jan 2015 – Jan 2015 | Greater Boston Area Built infrastructure in Scala and Spark to parallelize an energy-use prediction algorithm written in R, increasing prediction throughput by a factor of 1,000. This formed the backbone of a new product providing predictions to thousands of clients in real-time. ### Technology Associate Intern (Software Development) @ Bridgewater Associates Jan 2014 – Jan 2014 | Westport, CT Added an interface to a visualization application, enabling configuration of the application without writing code, saving dozens of hours of employee time per month. ### National Team Trainer @ Raffles Institution Jan 2013 – Jan 2013 - Trained national team that emerged as champions out of a field of 26 countries in the International Young Physicists Tournament. - Improved accuracy of team’s object recognition and tracking software via error detection and assisted correction. - Built a physics engine that simulated the effects of changes in experimental variables, allowing for more rapid discovery of novel phenomena. - Produced visualizations of experimental data in Mathematica to allow team members to effectively identify trends from a small series of test runs. Visualizations were also used in later research presentations. ### Intern @ Ministry of Trade and Industry Jan 2009 – Jan 2009 Prepared and presented report on how the government could better meet the economic aspirations of youth. ## Education ### Master of Engineering - MEng in Computer Science Massachusetts Institute of Technology Jan 2017 – Jan 2018 ### Bachelor of Science - BS in Electrical Engineering and Computer Science Massachusetts Institute of Technology Jan 2013 – Jan 2017 ### Raffles Institution Jan 2009 – Jan 2010 ### Raffles Institution Jan 2005 – Jan 2008 ## Contact & Social - LinkedIn: https://linkedin.com/in/vincent-tjeng --- Source: https://flows.cv/vincentt JSON Resume: https://flows.cv/vincentt/resume.json Last updated: 2026-03-23