Valentin Malykh
Head Of NLP Research at MTS AI
Moscow, Moscow City, Russia
Summary
🤩
Rockstar🎓
Top SchoolValentin Malykh is Head of NLP Research with 13 years of experience building and leading applied research teams in dialog systems, information retrieval, and large language models. He combines a strong academic foundation from MIPT and a PhD focused on noise robustness in NLP with hands-on engineering at companies like Huawei, VK, Yandex and MTS AI. His work spans research and production: from ConvAI backend fixes in the widely used ParlAI framework to full-stack implementations for content matching and subtitle processing in Huawei's Noah Research. Known for shipping practical solutions—news ranking, topic modeling, and QA/NLP4Code pipelines—he bridges deep learning research and scalable industry systems. Based in Moscow, he leads LLM research at MTS AI while maintaining active open-source contributions that reflect a pragmatic focus on dialogue interoperability and robustness.
13 years of coding experience
14 years of employment as a software developer
Deep Learning, Deep Learning at LxMLS 2016intern, distributed algorithms, intern, distributed algorithms at Laboratoire de Recherche en Informatique, CNRS, FranceMaster’s Degree, Computer and Information Sciences and Support Services, 4.7/5.0, Master’s Degree, Computer and Information Sciences and Support Services, 4.7/5.0 at Moscow Institute of Physics and Technology (State University) (MIPT)Doctor of Philosophy - PhD, Thesis: Noise Robustness in NLP Tasks, Doctor of Philosophy - PhD, Thesis: Noise Robustness in NLP Tasks at Russian State Certification
English, Russian, French
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (16)
machine-learning10
nlp10
sentence-transformers10
python10
natural-language-processing10
ml10
api-design9
scikit-learn9
scikit9
rest-api9
spacy9
beautifulsoup8
webscraping8
apidoc7
json7
Programming languages (6)
Github contributions (5)
huawei-noah/noah-research
May 2021 - May 2021
Noah Research
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:8 commits, 8 pushes in 1 day
Contributions summary:Valentin primarily contributed to the development of a system for processing and matching plot summaries and scenes, likely for a content analysis or recommendation system. This involved creating a `mapper.py` file with several classes and methods for feature extraction, sentence encoding, similarity scoring, and a dynamic time warping (DTW) algorithm. They also developed scripts for scraping movie data, converting data into a specific format, and processing subtitle files using Python.
noah
facebookresearch/ParlAI
Aug 2018 - Oct 2018
A framework for training and evaluating AI models on a variety of openly available dialogue datasets.
Role in this project:Backend Developer
Backend DeveloperContributions:11 commits, 8 PRs, 6 comments in 2 months
Contributions summary:Valentin primarily focused on modifying the backend logic of the ConvAI project. They addressed issues related to the ConvAI server output, modified URL formats, and increased the pull delay to handle server unavailability. Additionally, the user made adjustments to comply with the ConvAI2 interface and the PersonaChat format. The code changes involved modifications to the `convai_world.py` and `convai_bot.py` files.
nlpdeep-learningdatasetmachine-learningtraining
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.