Jacob Howcroft
Senior Software Engineer at Google
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolJacob Howcroft is a Senior Software Engineer in San Francisco with 12 years of experience building interactive experiences and production-quality software, currently shaping Google Doodles that blend play and machine learning. He brings deep hands-on expertise in both client and server domains—from iOS apps at Mango Languages to Django/Angular web work—and has contributed ML engineering improvements to Magenta’s Coconet model, optimizing TPU inference and SavedModel export workflows. At Google he’s shipped notable creative projects like the J.S. Bach and multiplayer Halloween Doodles, demonstrating an unusual mix of game development, ML integration, and production engineering. A University of Michigan computer science graduate, he pairs research-era prototyping experience with polished product delivery and a track record of refactoring systems for performance and reproducibility.
12 years of coding experience
2 years of employment as a software developer
Bachelor's Degree, Computer Science, Bachelor's Degree, Computer Science at University of MichiganOakland Community College
English, Spanish
Stackoverflow
305reputation
14kreached
5answers
4questions
Github Skills (15)
machine-learning10
tensorflow10
mlmodel10
save10
python10
data-model10
tpu9
objective-c6
xcode6
uicollectionviewlayout6
ios6
swift6
directory-structure6
uicollectionview6
overwrite6
Programming languages (4)
Github contributions (5)
magenta/magenta
Mar 2018 - Oct 2018
Magenta: Music and Art Generation with Machine Intelligence
Role in this project:
ML Engineer
ML EngineerContributions:13 commits, 17 PRs, 15 comments in 8 months
Contributions summary:Jacob primarily contributed to the Coconet model within the Magenta project, focusing on its implementation and optimization. Their work included adding documentation, refactoring code for TPU inference, and refactoring the code. They introduced utilities for exporting and loading SavedModels, including improvements to input and output data types. They also added features to override inputs during the sampling process.
artmusic-generationgenerative-artmagentamachine-learning
jhowcrof/magenta
Feb 2018 - Oct 2018
Magenta: Music and Art Generation with Machine Intelligence
Contributions:2 PRs, 35 pushes, 17 branches in 8 months
artmusic-generationmagentamachine-learningai
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.