Taras Sereda
Machine Learning Researcher at Gimlet Labs, Inc.
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolTaras Sereda is a Machine Learning Researcher based in San Francisco with 11 years of experience specializing in deep learning for speech and audio perception, particularly TTS, ASR, and source separation. He has led research teams and co-founded a startup focused on identity- and style-preserving multilingual speech synthesis, and has production experience deploying models on edge devices and GPUs. As a consultant he built real-time ASR for very low SNR audio and a zero-shot conversational TTS that runs faster-than-real-time and supports acoustic prompting (see collaboration with PolyAI). His background spans both research and hands-on engineering across startups and academia, including mentoring roles and lecturing on speech technologies. Notably, he combines practical ML operations—multi-GPU rigs and ARM deployments—with algorithmic work on diffusion, VAE, and neural synthesis methods to close the gap between SOTA research and deployable products.
11 years of coding experience
9 years of employment as a software developer
Polytechnical Lyceum of NTUU "KPI"Master's degree Mathematical Modelling, Master's degree Mathematical Modelling at Kyiv National Economics University
English, Ukrainian, Czech, Russian
Stackoverflow
420reputation
54kreached
4answers
1question
Github Skills (187)
python10
text-to-speech10
deep-learning10
speech10
gpu10
computer-graphics10
sequence-to-sequence10
ray10
gpu-acceleration10
cyclegan10
machine-translation10
multibox10
multimodal10
nmt10
numpy10
Programming languages (9)
Github contributions (5)
taras-sereda/cuda-machine-learning
Mar 2019 - Oct 2021
Machine Learning related CUDA code
Contributions:12 commits, 10 PRs, 58 pushes in 2 years 8 months
taras-sereda/audio-utils
Nov 2023 - Dec 2024
Contributions:3 PRs, 20 pushes, 3 branches in 1 year 1 month
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.