Peter Romov
PHD Student at Imperial College London
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolPeter Romov is a software and AI engineer with 12 years of experience combining senior technical roles, startups and academic research, currently a PhD student focused on privacy-preserving analytics. He has led teams and founded multiple ventures (GOSU, Data Souls, OpenDataScience) while building production ML systems and event platforms that scaled to tens of thousands of practitioners. At Meta and WhatsApp he specialized in differentially private ML, federated analytics and secure multi-party computation, contributing to Opacus and deploying privacy-by-design methods across major products. His open-source contributions include improving differential privacy support in PyTorch (Opacus) and enhancing fast topic modeling tooling (bigartm), reflecting both ML and backend craftsmanship. He blends hands-on engineering—shipping algorithms and APIs—with teaching and community leadership, having shaped ML education and competition formats adopted broadly. Unusually for an engineer of his profile, he holds patents and a track record of building privacy-preserving algorithms that passed regulatory review.
12 years of coding experience
13 years of employment as a software developer
Lomonosov Moscow State University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (16)
neural-network10
pytorch10
topic-modeling10
rnn-model10
machine-learning10
n10
python10
gru9
lstm9
pytorch-lightning8
deep-learning8
c-language8
deeplearning-ai8
documentation8
cprogramming-language8
Programming languages (7)
Github contributions (5)
bigartm/bigartm
Feb 2015 - Dec 2015
Fast topic modeling platform
Role in this project:
Back-end Developer
Back-end DeveloperContributions:61 commits, 9 PRs, 3 pushes in 9 months
Contributions summary:Peter primarily contributed to documentation and the Python API of the `bigartm` repository, a fast topic modeling platform. They added and updated documentation files, including footer templates and counters for analytics platforms like Yandex.Metrika and Google Analytics. Additionally, the user moved a Python directory outside the source and fixed installation issues. The user also fixed pep8 issues and travis script cleanup.
crosswordbigdatapythonldac-plus-plus
pytorch/opacus
Oct 2021 - Sep 2022
Training PyTorch models with differential privacy
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:58 reviews, 21 commits, 20 PRs in 10 months
Contributions summary:Peter implemented new features and refactored existing code within the `pytorch/opacus` repository, focusing on differential privacy for PyTorch models. Their work included adding support for GRU and RNN layers, refactoring LSTM implementations, and integrating the library with PyTorch Lightning for easier use. The user also fixed a bug in the DPOptimizer, contributing to the stability and usability of the library. Additionally, they updated the documentation and examples.
pytorchpytorch-modelsprivacydeep-learningprivacy-preserving-machine-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.