Oscar Knagg
Research Engineer at DeepMind
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolOscar Knagg is a research engineer based in London with 9 years’ experience applying physics-rooted intuition to data science and engineering problems across startups and research labs. He builds production-ready ML systems—from phonetic speech models and large-scale ETL to anomaly detection and Bayesian time-series—using Python and Scala, and deploying on AWS, Kubernetes, Spark and Docker. His favored tools reflect a probabilistic and deep-learning mix (convolutional networks, Gaussian processes, gradient-boosted trees), and he’s shipped solutions that saved hundreds of analyst-hours and enabled low-latency phishing detection at production precision. At DeepMind and G-Research he has bridged research and engineering, turning novel models into scalable pipelines. Outside work he documents experiments and tooling on GitHub and a technical blog, signaling a preference for building new things and explaining how they work.
10 years of coding experience
6 years of employment as a software developer
A Levels, A Levels, A Levels, A Levels at Bristol Grammar SchoolNatural Sciences (Exchange student), Physics, Natural Sciences (Exchange student), Physics at University of TorontoUniversity College London
English
Stackoverflow
148reputation
11kreached
2answers
1question
Github Skills (59)
pytorch10
convolutional-neural-networks10
language-model10
speaker-identification10
python10
meta-learning10
machine-learning10
maml10
numpy10
deep-learning10
gpu10
few-shot-learning10
neural-network10
autograd10
tensor10
Programming languages (7)
Github contributions (5)
oscarknagg/voicemap
Sep 2018 - May 2019
Identifying people from small audio fragments
Contributions:128 commits, 42 pushes, 2 branches in 8 months
audiomachine-learningspeaker-identificationspeaker-recognitionconvolutional-neural-networks
oscarknagg/wurm
Jan 2019 - Jul 2019
Massively multiagent reinforcement learning in a slither.io like environment
Contributions:197 commits, 6 PRs, 66 pushes in 5 months
multi-agent-reinforcement-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.