ABAB Pedia

Lasse Espeholt

Lasse Espeholt is a Danish machine-learning researcher and a co-founder of Ineffable Intelligence. He studied reinforcement learning at DeepMindGoogle DeepMindGoogle DeepMind is Google’s artificial intelligence research organization, formed in April 2023 by merging DeepMind, which Google acquired in 2014, with the Brain team from Google Research, and led by CEO Demis Hassabis. It developed AlphaGo, AlphaFold, and the Gemini family of models, and Hassabis and John Jumper received the 2024 Nobel Prize in Chemistry for AlphaFold’s protein structure prediction work.Open the full entry and Google BrainGoogle BrainGoogle Brain was Google’s AI research team, co-founded by Jeff Dean and others in 2011. It open-sourced the TensorFlow machine learning system in 2015 and was merged with DeepMind into Google DeepMind in April 2023.Open the full entry, helped start GoogleGoogleGoogle is an American technology company founded by Larry Page and Sergey Brin and incorporated in 1998. It became an Alphabet subsidiary in 2015. Its businesses include internet search, advertising, software platforms, cloud computing, and artificial intelligence.Open the full entry’s research team in Amsterdam and worked with colleagues on short-range precipitation forecasting.

Contents8 sections
Key facts

Childhood and education: From building websites to Aarhus University

Espeholt received his first computer at seven and played games on it. By ten he preferred programming websites for fun. At school in Denmark he especially liked analytical subjects. At Aarhus University he found that the whole computer science curriculum interested him and his results became more consistent. He studied for his bachelor’s degree from 2009 to 2012 and completed his master’s in the same field in 2014.⁠[1]

Consulting, LEGO and internships in the United States

While an undergraduate he ran his own consultancy and worked as a student developer at LEGO in 2011–2012. In 2012 he applied through Google’s website and spent fourteen weeks on developer tools in Seattle. He returned to Mountain View in 2013 and San Francisco in 2014, working on machine learning and cloud computing over the two placements. He negotiated with his university to reschedule classes around the internships. He later encouraged students seeking work abroad to prepare early and ask employees about referrals before applying.⁠[1]

DeepMind in London and distributed reinforcement learning

Google offered him a US job, but his visa application was unsuccessful. The company could place him in Zurich; he explored alternatives and joined DeepMind in London in January 2015 after seven interviews. He worked on machine translation, question answering and reinforcement learning, and began collaborating with David SilverDavid SilverDavid Silver is a computer scientist who led reinforcement learning research at DeepMind for many years, heading the AlphaGo, AlphaGo Zero and AlphaZero projects and co-leading AlphaStar. A professor at University College London and Fellow of the Royal Society, he left Google DeepMind in 2026 to focus on Ineffable Intelligence, the company he founded.Open the full entry around 2016. He was first author of the 2018 ICML paper on IMPALA, in which the team showed how one model could learn from experience collected across many tasks in parallel.⁠[1][2][3]

Starting a group in Amsterdam, forecasting weather and returning home

In the summer of 2018 he helped establish Google Brain’s Amsterdam group with a friend from DeepMind and an acquaintance from OpenAIOpenAIOpenAI is an artificial intelligence research and product organization founded in 2015 that develops the GPT model series, ChatGPT, Codex and a developer API, and conducts research on image, speech, video and intelligent agents. After its 2025 restructuring, the nonprofit OpenAI Foundation continues to control the for-profit OpenAI Group PBC. Its growth has involved large-scale fundraising and compute partnerships, and has also been accompanied by controversies over governance, copyright, privacy and AI safety.Open the full entry. He was expecting a child and considered Amsterdam a good place to raise a family. He worked with colleagues on precipitation forecasting. In 2022 he and his co-authors published a model for forecasts up to twelve hours ahead in Nature Communications. He later announced that he and his family had moved back to Aarhus and that he would continue working on weather with colleagues at Google Brain. Figurement’s team biography also records his work on large models and applied research at MicrosoftMicrosoftMicrosoft is a technology company founded by Bill Gates and Paul Allen in 1975 and headquartered in Redmond, Washington, with businesses spanning Windows, Microsoft 365, Azure, enterprise software, artificial intelligence, gaming and devices.Open the full entry. He also backed Figurement as an angel investor.⁠[1][4][5][6]

Joining the founding team at Ineffable Intelligence

In September 2026 Fortune identified Espeholt as one of six co-founders joining David Silver at Ineffable Intelligence. He later recalled that an investor had first contacted him to ask about Silver. The founding team then got in touch directly. It included former colleagues, and he welcomed the opportunity to pursue the work in Europe.⁠[7][3]

Learning and simulation with David Silver

In Sequoia CapitalSequoia CapitalSequoia Capital is a venture capital firm founded by Don Valentine in 1972, investing from the seed stage through the growth stage. Its historical investments include Apple, Cisco, Google, YouTube, Airbnb, and Stripe; in 2021 it introduced the Sequoia Capital Fund structure, which can continue to hold public securities, and in 2023 its regional businesses became independent brands and partnerships.Open the full entry’s interview he explained why he counted two departures from DeepMind. He left for Google Brain in 2018, became part of DeepMind again when the groups merged and subsequently left a second time. At around 39:11 he illustrated the difference between learning from simulations and observations with the Amsterdam team’s MetNet. Physical simulations accumulated error with time, he said, and the team tested learning short-term changes from observations for that particular forecasting task.⁠[3][4]

Related Organizations, Websites and Public Accounts

LEGO Group: Student developer. Official website: https://www.lego.com/ (opens in a new window).⁠[1]

DeepMind: Former researcher in London. Official website: https://deepmind.google/ (opens in a new window).⁠[1]

Google Brain: Helped start the Amsterdam research group, later part of Google DeepMind. Official website: https://deepmind.google/ (opens in a new window).⁠[1][3]

Microsoft: Worked on large models and applied research. Official website: https://www.microsoft.com/ (opens in a new window).⁠[6]

Ineffable Intelligence: Co-founder. Official website: https://www.ineffable.ai/ (opens in a new window).⁠[7][8]

Figurement: Angel investor. Official website: https://figurement.com/ (opens in a new window).⁠[6]

Sources9 sources

  1. Rising Star: Lasse var med til at starte Google’s AI afdeling i Amsterdam (opens in a new window)graduateships.com↑
  2. IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures (opens in a new window)proceedings.mlr.press↑
  3. David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience (opens in a new window)youtube.com↑
  4. Deep learning for twelve hour precipitation forecasts (opens in a new window)nature.com↑
  5. I moved back to Aarhus, Denmark (opens in a new window)linkedin.com↑
  6. Figurement Team: Lasse Espeholt (opens in a new window)figurement.com↑
  7. Ineffable Intelligence adds six cofounders (opens in a new window)fortune.com↑
  8. Ineffable Intelligence (opens in a new window)ineffable.ai↑
  9. Lasse Espeholt (lespeholt) (opens in a new window)github.com↑