Kevin Jose
Member Of Technicall Staff at Reflection AI
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolKevin Jose is a machine learning generalist with 12 years of experience applying research-grade models to production problems, currently focused on ML at Reflection AI after recent roles at Meta and Amazon. He has hands-on expertise across large-scale model development and deployment—working on foundation models, ablation studies, and massive Spark optimizations for web-scale data—as well as earlier work on language models for Alexa and synthetic data/knowledge distillation. His background spans research and engineering roles from Max Planck Institute contributions to open-source HTM examples in the notable Numenta nupic-legacy repo, showing an appetite for neuroscience-inspired approaches as well as pragmatic systems. Based in London, he pairs a strong academic foundation (MS in CS) with decades of backend, DevOps, and applied-science experience that lets him bridge experiment and production effectively.
12 years of coding experience
9 years of employment as a software developer
Bachelor of Technology (BTech), Computer Science, Bachelor of Technology (BTech), Computer Science at College of Engineering TrivandrumMasters, Computer Science, Masters, Computer Science at Universität des Saarlandes
English, Malayalam, German
Stackoverflow
640reputation
75kreached
16answers
23questions
Github Skills (12)
hierarchical-temporal-memory10
artificial-intelligence10
python10
numpy9
flood-fill6
react-router6
javascript6
game-physics6
game-engine6
development-environment6
django6
react6
Programming languages (9)
Github contributions (5)
numenta/nupic-legacy
Mar 2014 - Mar 2014
Numenta Platform for Intelligent Computing is an implementation of Hierarchical Temporal Memory (HTM), a theory of intelligence based strictly on the neuroscience of the neocortex.
Role in this project:
ML Engineer
ML EngineerContributions:5 commits in 10 days
Contributions summary:Kevin contributed to an example within the repository demonstrating the use of the Spatial Pooler (SP) component, a core element of Hierarchical Temporal Memory (HTM) architectures. The user implemented a program to illustrate the SP's functionality, including input vector creation, execution, and noise addition. The commits demonstrate an understanding of the SP's parameters and its behavior with varying inputs.
htmmemoryneurosciencedeep-learninghierarchical
kevinmartinjos/lottery-ticket-experiments
Jun 2019 - Aug 2019
Contributions:81 commits, 77 pushes, 1 branch in 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.