Elena Rastorgueva

I am an applied researcher at NVIDIA, specializing in speech AI.

I grew up in London and have been living in the San Francisco Bay Area since 2022.

Work

I work on the NVIDIA AI Applications team. We develop new speech models such as Nemotron Voicechat and Parakeet. Our work feeds into NVIDIA NeMo and NVIDIA Riva.
Selected talks: NeMo Framework talk at a Linux Foundation event; "Speech AI Demystified" talks at GTC 2024 and GTC 2025.
Selected blogs: How forced alignment works - explainer of Viterbi decoding, with animations.
Before NVIDIA, I interned at Amazon Alexa, working on Text To Speech, and at Mila, working on the Speechbrain toolkit.
I became interested in speech after working on Mandarin Tone Trainer during college. It's a website that helps you practice speaking Mandarin. It reached #1 on Hacker News when I released it.

I studied Engineering at the University of Cambridge, specializing in Information & Computer Engineering. My MEng thesis was on applying deep learning to traditional machine translation techniques.
While at Cambridge, I was on the Cambridge University Eco Racing solar car team and drove the car in the European Solar Challenge.

In my spare time I study Mandarin. Most recently I passed the B1 level of the Test of Chinese as a Foreign Language.
I also like learning how computers work from the ground up by reading textbooks and taking courses such as Nand2Tetris.