Piotr Dabkowski
InvestorUnicorn founderFounder at Eleven Labs · Co-founder at ElevenLabs
United States
FoundedEleven Labs$11BElevenLabs
Find your path to Piotr
432 people in our graph share verified history with Piotr — schools, employers, funds. One of them is your warm intro.
About
Before co-founding ElevenLabs in 2022, he spent four years as a software engineer at Google, following earlier engineering work at Tessian. He also completed internships at Google and Opera Software.
Likely email· guessed from name + firm domain, not firm-published
- p•••@elevenlabs.ioreveal75% confident
Invests in
AI & Machine Learning
Beyond investing
- Founded
- After the cofounders found incumbent speech systems insufficiently scalable, controllable, or human, Dąbkowski built the initial text-to-speech models from scratch. The key early result inferred context to produce more natural delivery; this attribution comes from Staniszewski's account published by an ElevenLabs investor.
- Awards
- Named with Staniszewski to Forbes' 2024 Europe Technology list for ElevenLabs. Selected for TIME's 2024 list of influential figures in artificial intelligence, with a profile focused on ElevenLabs' model quality, multilingual dubbing, provenance tools, and remaining misuse risks.
- More
- With Cambridge supervisor Yarin Gal, published a NeurIPS 2017 method that learns sharp image-saliency masks in one forward pass. The paper reports about 100 masks per second, outperforming other weakly supervised methods on ImageNet localization, and introduces a saliency-quality metric. Began Js2Py in 2014 as a pure-Python JavaScript translator and interpreter, later also publishing pyjsparser and the reference implementation for his saliency research. Public repository and package records identify him as the principal author. ElevenLabs launched its public beta in January 2023 with text-to-speech, voice cloning, and voice design. Dąbkowski framed the long-term goal as letting people enjoy relevant content regardless of language. Coauthored the launch of Eleven v3 alpha, built to add audio tags, multi-speaker dialogue, interruptions, and deeper expressive control across more than 70 languages; the authors explicitly disclosed its alpha-stage reliability and latency limits. In ElevenLabs' $500 million Series D announcement at an $11 billion valuation, Dąbkowski described the research remit as foundational models spanning text-to-speech, transcription, music, dubbing, and conversation, optimized together with products. Financing and valuation are company-reported, not personal liquidity. Met future cofounder Mati Staniszewski at about age fifteen in the same IB class at Warsaw's Kopernik 33rd high school. They bonded over mathematics and later studied, traveled, lived, and built side projects together. At a rare 2025 live interview, an attendee quoted Dąbkowski saying he had wanted to attend Oxford since age nine and thanking his father for teaching him mathematics and physics. The recording was announced but no complete, independently reviewable transcript was found, so this remains attributed to the attendee report.
From Piotr's own website — common ground for a warm intro.
Founder at Eleven Labs
Intro paths
In turn, Piotr can intro founders to Eleven Labs's investors below.
Eleven Labs's investors · 6
Co-founders at Eleven Labs
Co-founder at ElevenLabs
Intro paths
In turn, Piotr can intro founders to ElevenLabs's investors below.
ElevenLabs's investors · 31
Co-founders at ElevenLabs
Experience
- ElevenLabsCo-Foundercurrent2022 — present
- GoogleSoftware Engineer2018 — 2022Zürich Area, Switzerland
- TessianSoftware Engineer2017 — 2018London, United Kingdom
- GoogleSoftware Engineering Intern2016Warsaw, Masovian District, Poland
- Opera Software ASSoftware Engineering Intern2015 — 2016Oslo Area, Norway
Investments
Investor
BlandElastics
Per curated public sources.
Founder fit — who Piotr backs
Potentially relevant to research-led AI founders in audio, conversational systems, developer infrastructure, or quantitative applications; this is a cautious pattern observation, not stated criteria.
[]
How to pitch
No public pitching instructions were found. His 2025 founder advice, reported by an interview attendee, emphasized a strong support network of angels, mentors, and people with soft power rather than a sector-specific pitch formula.
Sectors: Artificial intelligence, Voice and conversational AI, Prediction-market and quantitative tooling
Editorial inference from Piotr's portfolio, writing and public record — not a published mandate.
Investment themes
5 documented themes from Piotr's essays, talks and portfolio, grouped by what they say about founders.
▶More themes5
Research-led, context-aware audio2022
His remit centers on models that understand linguistic and conversational context rather than merely rendering clean speech, with expressiveness, emotion, pacing, and interruptions treated as core model capabilities.
For founders: Relevant to founders whose differentiation depends on original model research rather than a thin application layer.
Models optimized with products2026
Publicly argues that foundational models should be optimized for the product experiences they enable, across text-to-speech, transcription, dubbing, music, and conversational systems.
For founders: Signals interest in teams that can translate research advantages into measurable user experiences.
Language access2023
Frames the company's long-term objective as making relevant spoken content available regardless of the listener's language, rooted in the founders' experience of low-quality Polish voiceover.
For founders: Relevant to accessibility, localization, media, education, and cross-language infrastructure.
Traceability and controlled access2024
As model quality approaches human indistinguishability, he emphasizes traceable generations, stronger access checks for voice cloning, and synthetic-audio classification. He told TIME that open-source audio models pose a harder control problem, while TIME noted that safeguards remain imperfect and evolving.
For founders: Relevant to founders building provenance, identity, moderation, detection, and trustworthy deployment infrastructure.
Fast, practical interpretability2017
His Cambridge work favored a model-agnostic explanation method that produced useful image saliency in one forward pass, reflecting an early pattern of pairing research novelty with deployable speed.
For founders: Relevant to explainability, evaluation, computer vision, and production ML tooling.
Podcasts & interviews
featuring Piotr DabkowskiFirst public live interview at This is it
Mark Bain on LinkedIn · linkedin.com
Piotr Dabkowski: The 100 Most Influential People in AI 2024
TIME · time.com
Meet the Future of AI Voice — The ElevenLabs Documentary
YouTube
Piotr Dabkowski
Forbes · forbes.com
Writing
Cofounder-authored announcement outlining the expansion from human-like voice into foundational models across the audio stack and the importance of optimizing research for products.
Cofounder-authored model launch explaining expressiveness, audio tags, dialogue mode, multilingual coverage, and candid alpha limitations.
Public PyTorch implementation and real-time webcam demonstration accompanying the NeurIPS saliency paper.
Peer-reviewed NeurIPS paper with Yarin Gal presenting a fast learned masking model, a new saliency metric, and strong ImageNet localization results.
Long-running MIT-licensed JavaScript-to-Python translator and interpreter for which package and repository records identify Dąbkowski as author and principal contributor.
Frequently asked questions
- What is the canonical spelling of Piotr Dabkowski's name?
- Current ElevenLabs materials use Piotr Dąbkowski, preserving the Polish ą. Academic, code, professional, and UK registry records commonly render it as Piotr Dabkowski; both forms refer to the same person.
- What is Piotr Dąbkowski's role at ElevenLabs?
- He is cofounder and is publicly described as CTO or research head, leading research and engineering teams across ElevenLabs' audio models. Forbes notes that the company historically avoided rigid titles, so CTO is best treated as the public functional title rather than evidence of a conventional hierarchy.
- What did Piotr Dąbkowski do before ElevenLabs?
- He studied engineering at Oxford, completed an Advanced Computer Science MPhil at Cambridge, published at NeurIPS, interned at Opera and Google, briefly worked at Tessian, and then spent March 2018 to January 2022 at Google working in software engineering and machine learning.
- What was Piotr Dąbkowski's academic research?
- His Cambridge work with Yarin Gal trained a masking model to explain differentiable image classifiers in a single forward pass. Published at NeurIPS 2017, it reported real-time performance, strong ImageNet localization, and a new way to score saliency explanations.
- How did Piotr Dąbkowski and Mati Staniszewski start ElevenLabs?
- The friends met in a Warsaw IB class around age fifteen and kept building side projects while Dąbkowski worked at Google and Staniszewski at Palantir. Their dissatisfaction with one-narrator Polish dubbing and weak existing speech systems led them to pool savings, leave their jobs, and build a context-aware text-to-speech model in 2022; the public beta followed in January 2023.
- What is his public position on synthetic-audio safety?
- He supports traceability of every generation, identity and payment friction around voice cloning, and classifiers for ElevenLabs-generated audio. He has also said open-source audio models are harder to control; TIME cautions that circumvention and the changing threat landscape mean no safeguard is complete.
- Does Piotr Dąbkowski invest personally?
- Yes, but the verified public record is small. Elastics' founders and reputable round reporting identify him as a pre-seed angel, while Bland's Series C announcement names him as an existing investor who participated again. No reliable public source reviewed gives his check sizes or a first-person thesis, and ElevenLabs' own backers are not treated as his investments.
- What boards or advisory roles does Piotr Dąbkowski hold?
- UK Companies House verifies one active appointment: director of ELEVEN LABS LTD since January 4, 2022, where he is also recorded as a person with significant control. No reliable public source reviewed established an outside board or advisory role.
- Who is Piotr Dabkowski?
- Before co-founding ElevenLabs in 2022, he spent four years as a software engineer at Google, following earlier engineering work at Tessian. He also completed internships at Google and Opera Software.
- Where does Piotr Dabkowski work?
- Piotr Dabkowski is Founder of Eleven Labs and Co-founder at ElevenLabs.
- What does Piotr Dabkowski invest in?
- Piotr Dabkowski invests in AI & Machine Learning.
- What companies did Piotr Dabkowski found?
- Piotr Dabkowski founded Eleven Labs, ElevenLabs.
- What companies has Piotr Dabkowski invested in?
- Investments tracked here include Bland, Elastics.
- Where did Piotr Dabkowski go to school?
- Piotr Dabkowski studied at University of Cambridge (Master of Philosophy - MPhil, Advanced Computer Science); University of Oxford (Bachelor of Arts - BA, Engineering).
Founder mafias
University of Oxford mafia623
Also came through, then founded or led a company.
+ 616 more
University of Cambridge mafia588
Also came through, then founded or led a company.
+ 581 more
Tessian mafia10
Also came through, then founded or led a company.
+ 3 more
Same school
Also attended University of Cambridge1,412
.png)
+ 1,406 more
Also attended University of Oxford1,553
+ 1,547 more
Shared employers
Also worked at ElevenLabs38
.png)
.png)
+ 32 more
Also worked at Google2,000
+ 1,994 more
Also worked at Tessian19
+ 13 more
Also worked at Opera Software AS12
+ 6 more