Applied speech AI

Research that ships.

The products run live on this site; the benchmarks behind them are published in the open, and some of the research is too.

The story

It started with a question.

The question

Could speech itself adapt to the person listening, live, and still sound like the one speaking?

What we saw

We came from speech research at Google, where the work was teaching machines to hear. The person listening was still doing the hard part: the learner replaying the podcast, the audience at the session nobody could afford to interpret, the patient nodding along to the doctor. The speaker, however willing, cannot help each of them at once.

What we set out to build

Speech that meets every listener: interpreted into their language, said again at their level, kept in the speaker's own voice, and run wherever the audio is allowed to be. We started with the live case, because it is the hardest, and we put the results on this site as they happen.

On air →

People

Meet the team.

CEOFounder

Dominik Roblek

Speech and audio ML at Google Research for a decade; co-author of Now Playing, Fréchet Audio Distance, SEANet, SPICE, and AudioLM.

LinkedIn X GitHub

CTOFounder

Hassan Rom

Fifteen years of applied ML at Google: Search, Shopping, perception for Waymo and Image Search, and on-device audio understanding for Android, Pixel, and Nest. Co-creator of the Open Images dataset.

LinkedIn

UI/UXFounding member

Ammar Raji

Designing and building intuitive experiences for real-time speech AI products, from early concepts to production.

LinkedIn

ApplicationsFounding member

Amwalul Ikhsani

Software engineer specializing in AI-powered applications and scalable products.

LinkedIn

ML Research

Mariam Elghandoor

ML Research Engineer working on real-time speech and NLP. Zewail City AI & Data Science graduate with multiple first-place hackathon wins; prior ML work spans conversational AI, LLMs, and on-device models.

LinkedIn

ML ResearchFounding member

Muhammad Essam

ML Research Engineer, building real-time speech models. Zewail City Communications and Information Engineering graduate.

LinkedIn

Applied MLFounding member

Yahya Saleh

Applied ML Engineer at VoiceFrom: real-time speech AI, on-device models, from papers to production.

LinkedIn GitHub

There's room at this table

Take a seat.

Talk to us.

Tell us what you're planning. We reply from hello@voicefrom.ai.

Email us instead