Luke De Oliveira
Head Of Artificial Intelligence at Rokos Capital Management
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolLuke De Oliveira is a seasoned AI leader overseeing AI strategy and delivery at Rokos Capital Management in London, bringing more than a decade of hands-on ML engineering and leadership. He previously led AI research and engineering at Meta on distillation, inference, and pruning to accelerate large language models, and steered the AI organization at Twilio across NAMER and EMEA, including Voice Intelligence and transformer-based dialogue applications. As co-founder and former CEO of Vai Technologies (acquired by Twilio), he built few-shot learning and adaptive curricula for language understanding with minimal data. His early research spans CERN, SLAC, and Yale Law School, focusing on high-energy physics and agent-based simulations, culminating in publications and contributions to deep generative models for physics and social-spatial simulations. An active open-source contributor, he has worked on keras-team/keras, implementing training interruption handling, new layers, and examples like ACGAN on MNIST, reflecting a hands-on, production-focused ML mindset. Luke holds a BS in Applied Mathematics with Economics & ML from Yale and an MS in Computational and Applied Mathematics from Stanford, blending rigorous math with practical AI system design.
12 years of coding experience
11 years of employment as a software developer
Bachelor of Science (B.S.), Applied Mathematics, concentrations in Economics & Machine Learning, Bachelor of Science (B.S.), Applied Mathematics, concentrations in Economics & Machine Learning at Yale UniversityMaster of Science (M.S.), Computational and Applied Mathematics, Master of Science (M.S.), Computational and Applied Mathematics at Stanford University
French, English
Github Skills (7)
keras10
tensorflow210
machine-learning10
deep-learning10
tensorflow10
neural-networks10
python10
Programming languages (6)
Github contributions (5)
keras-team/keras
May 2015 - Aug 2017
Deep Learning for humans
Role in this project:
ML Engineer
ML EngineerContributions:15 commits, 11 PRs, 83 comments in 2 years 3 months
Contributions summary:Luke primarily focused on modifying the `keras/models.py` file to handle keyboard interrupts during training and incorporated features. They also made changes to `keras/layers/core.py`, and integrated `MaxoutDense` layer. The user further merged and updated core files, including code associated with highway networks and an example for an ACGAN using MNIST data. This indicates a focus on model development and feature enhancements within the Keras deep learning framework.
pythondata-sciencedeep-learningneural-networksmachine-learning
lukedeo/torch-serving
Aug 2019 - Oct 2020
Contributions:1 release, 42 commits, 39 pushes in 1 year 2 months
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.