Kittipat Virochsiri
Software Engineer Machine Learning at Facebook
California, United States
Summary
🤩
Rockstar🎓
Top SchoolKittipat Virochsiri is a Software Engineer specializing in machine learning with 12 years’ experience building large-scale, production ML and backend systems from California. At Facebook he applied reinforcement learning and real-time indexing to improve search and ranking across diverse entity types, and at Instagram he led backend work for Explore relevance and Stories integration. He’s an experienced contributor to notable open-source projects such as Caffe2 and ReAgent, where he implemented core operator changes, sampling mechanisms, and parametric DQN components to boost training efficiency and model expressiveness. Comfortable across feature engineering, retrieval, and ranking pipelines, he blends deep ML model work with pragmatic engineering for production performance. His academic background from RWTH Aachen and Edinburgh underpins a research-minded approach to scalable systems and applied RL.
12 years of coding experience
2 years of employment as a software developer
MSc, Informatics, MSc, Informatics at The University of EdinburghMSc, Informatics, MSc, Informatics at RWTH Aachen UniversityBachelor of Engineering, Computer Engineering, Bachelor of Engineering, Computer Engineering at Chulalongkorn University
English, Thai
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (16)
dqn10
machine-learning10
pytorch10
caffe10
deep-learning10
tensorflow10
python10
reinforcement-learning10
testing10
back-end-development9
pytorch-lightning8
data-structure7
data-structures7
algorithm7
algorithms7
Programming languages (11)
Github contributions (5)
facebookresearch/ReAgent
Aug 2018 - Sep 2021
A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1 release, 358 commits, 135 PRs in 3 years 1 month
Contributions summary:Kittipat contributed to the development of a platform for reasoning systems, including reinforcement learning and contextual bandits. Their work involved implementing and testing parametric DQN with a factorization Q-function, which suggests the user was involved in designing and implementing machine learning models. The user also contributed to testing and debugging code related to the gridworld environments, which further implies development focus. Code changes related to the structure of the model's components indicate back-end development skills.
reinforcement-learningcontextualbanditscontextual-banditsreinforcement
facebookarchive/caffe2
Feb 2017 - Mar 2018
Caffe2 is a lightweight, modular, and scalable deep learning framework.
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:136 commits in 1 year 1 month
Contributions summary:Kittipat primarily contributed to the Caffe2 deep learning framework by modifying core operator schemas and implementing new functionalities. Their work involved changes to index operators, specifically regarding state management, and the creation of a unique uniform fill operator. Further contributions included the addition of a sampling mechanism for softmax, and the development of supporting layers related to sampling and the use of ID-score list features, indicating a focus on improving training efficiency and supporting new data formats for machine learning workflows.
pytorchscalablecaffe2deep-learningml
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.