Aditya Singh
Senior Machine Learning Engineer at Sky
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolAditya Singh is a Senior Machine Learning Engineer based in London with 12 years of experience building computer vision and ML systems across industry and research. He began with MS research on short-form video understanding—action recognition and hashtag recommendation—and has since shipped solutions for video ad exchanges and analytics as a platform and CV engineer. His roles span startups and enterprises (Directi, Verisk, Cortexica, Zebra, Sorted and now Sky), blending applied research with production engineering to take models from prototype to scale. An active contributor to JAX-based tooling, he implemented a suite of adaptive pooling layers for the popular Equinox library, signaling deep familiarity with neural network primitives as well as practical deployment constraints. Not obvious from titles: he often sits at the intersection of advertising systems and visual analytics, making him adept at aligning ML research with business KPIs.
12 years of coding experience
9 years of employment as a software developer
Bachelor's degree Computer Science, Bachelor's degree Computer Science at International Institute of Information Technology Hyderabad (IIITH)
Github Skills (8)
neural-network10
deep-learning10
jax10
python10
algorithms8
data-structures8
algorithm8
data-structure8
Programming languages (4)
Github contributions (5)
patrick-kidger/equinox
Jul 2022 - Aug 2022
Elegant easy-to-use neural networks + scientific computing in JAX. https://docs.kidger.site/equinox/
Role in this project:
ML Engineer
ML EngineerContributions:19 reviews, 28 commits, 7 PRs in 1 month
Contributions summary:Aditya primarily contributed to the implementation of adaptive pooling layers within the Equinox library, a JAX-based neural network framework. Their work involved adding `AdaptiveAvgPool1d`, `AdaptiveAvgPool2d`, and `AdaptiveAvgPool3d` with associated test cases. Additionally, they introduced `AdaptiveMaxPool1d`, `AdaptiveMaxPool2d`, and `AdaptiveMaxPool3d`, providing a comprehensive suite of adaptive pooling functionalities. These contributions demonstrate a focus on extending the library's capabilities for deep learning model development and potentially improving model performance through adaptive pooling techniques.
pytorchdeep-learningneural-networkscallablejax
paganpasta/eqxvision
Jul 2022 - Jan 2023
A Python package of computer vision models for the Equinox ecosystem.
Contributions:24 releases, 1 review, 200 commits in 5 months
pytorchpythonvisioncomputer-visionimage-processing
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.