Oktai Tatanov
Russia
Summary
🤩
Rockstar🎓
Top SchoolOktai Tatanov is a machine learning engineer with a decade of experience and over five years focused on ML and deep learning research in both academic and industrial settings. He has shipped speech synthesis and TTS improvements at scale—contributing to NVIDIA's NeMo TalkNet, HiFiGAN and accompanying TTS pipelines—and worked on molecular generation benchmarks in the MOSES project. His roles span startups and industry leaders (Neuromation, VK, NVIDIA, Play.ht, Rask AI), giving him a practical track record of turning research prototypes into production-ready models. Based in Russia and trained at ITMO University, he combines hands-on model engineering with attention to training workflows, data aligners, and G2P components—skills that often sit between research and production.
10 years of coding experience
5 years of employment as a software developer
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at ITMO University
Stackoverflow
123reputation
4kreached
2answers
4questions
Github Skills (20)
pytorch10
drug-discovery10
python10
machine-learning10
text-to-speech10
rnn-model10
large-language-models10
n10
deep-learning10
generative-ai10
neural-network10
generative-model10
benchmark9
benchmarking9
pandas8
Programming languages (8)
Github contributions (5)
molecularsets/moses
Jun 2018 - Nov 2018
Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models
Role in this project:
ML Engineer
ML EngineerContributions:11 commits, 1 PR in 5 months
Contributions summary:Oktai primarily contributed to the development and modification of molecular generation models within the MOSES repository, focusing on techniques such as character-level RNNs and junction trees. Their work involved implementing and refining model architectures, as evidenced by code changes in model definitions and training scripts. These changes involved adjusting metrics, particularly with the inclusion of FCD, along with integrating character RNNs and junction trees.
cheminformaticsbenchmarkingchemistrymolecular-generationdrug-discovery
NVIDIA/NeMo
Apr 2021 - Feb 2022
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
Role in this project:ML Engineer
ML EngineerContributions:110 reviews, 53 commits, 58 PRs in 10 months
Contributions summary:Oktai primarily contributed to the development and enhancement of the TalkNet model within the NeMo framework. Their work involved bug fixes, the addition of a TalkNet training tutorial, and improvements related to the TTS aligner and G2P components. Code changes also included modifications to HiFiGAN, suggesting involvement in the complete TTS pipeline. These contributions focused on improving model functionality, training procedures, and overall code quality within the speech synthesis domain.
asrspeech-recognitionnatural-language-processingttsspeaker-diarization
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.