Vlad Grigorev - Machine Learning Engineer, Text-to-Speech

Moscow, Russian Federation

Experience

Dec 2021 - Present

3 years 8 months

Machine Learning Engineer, Text-to-Speech

Tensor

Implemented Text-to-Speech model into production on AWS-like cloud with Docker and Kubernetes.
Developed baseline for NLP text processor, the testing pyramid and monitoring. Text-to-Speech service is now adopted by wider CRM and Tech Support teams with 10k of end clients.
Designed and developed NLP text normalisation baseline, microservice structure and a framework for the Text-to-Speech synthesis. That speeds up implementation of new ML models reducing time to market to weeks.
Enhanced CI/CD and built MLOps pipeline. That cut down implementation time by 8 times and enabled outstandingly fast and seamless scaling.

Jun 2019 - Dec 2021

2 years 7 months

Computational Physics Research Engineer

Joint Institute For High Temperatures

Developed a numerical algorithm for shock wave computation and adopted a semiclassical toolkit to take into account quantum effects. That significantly improved accuracy yet the processing time is comparable to the classic model. The new algorithm accelerates the computation of the state of matter under a given set of physical conditions up to 10-80 times, providing a fast and convenient starting point to research electrical and magnetic properties.
Technologies: Python, Pytorch, Jupyter, Pandas, Numpy, Scipy, Matplotlib, C/C++, Git, Linux.

Summary

A young professional deeply immersed into Speech Technologies. Having background in Academia, I am looking for the MLE position to contribute to real-world applications and bring a new set of capabilities to the team.