Andrei Kapishnikov
Software Engineer at Google DeepMind
Boston, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolAndrei Kapishnikov is a software engineer with 11 years of experience (and a decade-plus career) building large-scale, fault-tolerant systems and cutting-edge ML research products from infrastructure to explainability. Currently at Google DeepMind, he led development of non-autoregressive generative algorithms, created explainability methods (XRAI, Guided IG) integrated into Google Cloud and cited in Nature and PNAS, and shipped high-impact services used by billions such as Cronet. His background includes core roles on VMware vSAN and Oracle Entitlements Server, reflecting deep expertise in distributed storage, security, and device management. He contributes to robust ML tooling and QA—adding unit tests and enhancements to the well-known PAIR-code saliency framework—bridging research and production engineering. Trained with first-class honors (MSc, University of Latvia), he pairs rigorous academic grounding with a track record of publishing, patenting, and productizing research. Not obvious at first glance: he combines explainability research with practical deployment experience, making him effective at turning novel algorithms into scalable, production-ready services.
11 years of coding experience
13 years of employment as a software developer
Master of Science (M.Sc.) | First-class honors (GPA 9.6/10) Computer Science, Master of Science (M.Sc.) | First-class honors (GPA 9.6/10) Computer Science at University of Latvia
Github Skills (7)
unit-testing10
tensorflow10
python10
machine-learning9
deep-learning9
convolutional-neural-networks8
image-recognition8
Programming languages (2)
Github contributions (5)
PAIR-code/saliency
Aug 2019 - May 2022
Framework-agnostic implementation for state-of-the-art saliency methods (XRAI, BlurIG, SmoothGrad, and more).
Role in this project:
QA Engineer / Test Automation Engineer
QA Engineer / Test Automation EngineerContributions:38 reviews, 9 commits, 23 PRs in 2 years 9 months
Contributions summary:Andrei primarily focused on implementing and expanding unit tests for the XRAI attribution algorithm within the saliency framework. They added new test cases to verify different configurations of XRAI, including full and fast algorithms with and without segment flattening, as well as custom segments and baselines. Additionally, the user made adjustments to the core test to ensure the patch correctly stops. A core contribution was also made allowing for pre-calculated base attribution to XRAI.
pytorchartimage-recognitionsaliency-mapdeep-learning
gkapish/saliency
May 2019 - May 2022
TensorFlow implementation for SmoothGrad, Grad-CAM, Guided backprop, Integrated Gradients and other saliency techniques
Contributions:4 PRs, 22 pushes, 19 branches in 3 years
techniquesgradientsintegratedintegrated-gradientsdeep-learning
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.