What is the DeepMind mafia?

According to Business Insider, the DeepMind mafia are a team of 18 alumni from Google’s AI research lab who have raised millions of dollars for their own start-ups and non-profits. 

Karl Moritz Hermann

Karl Moritz Hermann is a research scientist who spent the back end of his five-year stint at DeepMind as a senior researcher at its Berlin office. He joined OpenAI after the company acquired his data and NLP start-up Dark Blue Labs in 2014. 

Before this role, Hermann completed postdoctoral research in machine learning and computational linguistics at Oxford University. 

He then co-founded another start-up called Saiga with Sophia Höfling and Michael Fialik in Berlin. Saiga is developing an AI-powered personal assistant under the same name which Hermann notes frees up time by handling everyday admin tasks such as “scheduling dentist appointments, or booking restaurants and holidays.

Trevor Beck

Trevor Beck worked in various product management roles at DeepMind for over nine years. In 2015, he launched the company’s mobile clinical alert app and then led its healthcare research team between 2016 and 2018. He also worked for a time on the commercialization of DeepMind’s artificial intelligence.

Beck is the founder and current CEO of Shift Lab, a company that builds impactful artificial intelligence. Genome, the company’s first product, delivers an eCommerce and marketing solution to sellers and creators that uses generative AI and language models.

Mustafa Suleyman

Mustafa Suleyman co-founded DeepMind Technologies with Shane Legg and Demis Hassabis. For almost a decade, he worked as Chief Product Officer and then as Head of Applied AI before leaving the company in early 2020. However, he continued at DeepMind owner Google for another two years as its VP of AI Product Management & AI Policy.

In February 2022, Suleyman founded Inflection AI with LinkedIn co-founder Reid Hoffman and former colleague Karén Simonyan. Little is known about the company to date, but CNBC’s Sam Shead claims Inflection AI wants to “develop AI software products that make it easier for humans to communicate with computers.

Suleyman is also a venture partner at VC firm Greylock Partners.

Andrew Eland

Andrew Eland is a Google veteran who worked for the company for over 11 years. He primarily worked on Google Maps and was involved in local business discovery, international expansion, location data infrastructure, city mobility data, and maps for crisis response, elections, flu trends, data philanthropy, and civic engagement.

At DeepMind, Eland was responsible for the engineering behind DeepMind Health (now Google Health) to tackle some of the industry’s most complex problems. However, he quit his senior role to found the city planning consultancy firm Diagonal in 2019.

Eland started the company after observing that no other company offered the type of urban data analysis software he envisaged. To that end, Diagonal works with planners, real estate developers, public and research institutions, and urban design studios to address problems with large or complex datasets. These data may describe any information relevant to cities such as environmental factors, transport networks, demographics, economics, and buildings.

Adji Bousso Dieng

Adji Bousso Dieng is a Senegalese statistician and computer scientist who researches deep learning and probabilistic models to derive meaningful structures from unlabelled data. Dieng was an intern at DeepMind in London and also spent time at Microsoft Research and Facebook AI Research.

Dieng joined Google Brain as a research scientist in August 2020 and continues to serve in the position today. She is also the founder of The Africa I Know, a non-profit that seeks to shift the narrative around Africa and equip young African students with the educational support they need to drive the future development of the continent. 

Peter van Toth

Peter van Toth is a former AI researcher with extensive experience in the design and implementation of solutions that solve ML and numerical optimization problems. 

In November 2021, van Toth announced $5 million in seed funding for a new venture that would enable web3 to benefit from advanced AI. He started a company known as Kosen Labs with a team of former DeepMind colleagues such as Miljan Martic. 

After many years spent working on AI techniques for closed-source systems, van Toth believed it was time to make the benefits of AI systems open and permissionless. One of the first projects Kosen Labs tackled was the DeFi issue of active liquidity management on modern, decentralized exchanges.

Tiago Ramalho

Tiago Ramalho worked at DeepMind for over three years and was eventually employed as a Research Engineering Manager. Ramalho is also a physicist who has published academic papers on neural networks.

In September 2018, Ramalho left DeepMind to work as a research scientist at the Japanese AI company Cogent Labs. Two years later, he founded Recursive AI – a start-up that builds AI models to help companies increase efficiency and better manage their sustainability endeavors. 

Hugo Penedones

Hugo Penedones is a machine learning researcher and engineer who worked at DeepMind in both London and Zurich. Penedones is a founding member of DeepMind’s AlphaFold project which can predict 3D models of protein structures to prevent disease, find new medicines, and advance nearly every field in biology.

Penedones then went on to co-found Inductiva Research Labs where machine learning is being applied to computational problems in mathematics, chemistry, biology, and physics. He currently serves as Inductiva’s CTO.

Martin Schmid, Matej Moravcik, and Rudolf Kadlec

Martin Schmid is an AI research scientist who spent almost a decade at IBM and then DeepMind. He later moved to the Czech Republic and founded EquiLibre Technologies with former co-workers Matej Moravcik and Rudolf Kadlec.

EquiLibre is a company that applies AI to the trading industry with aspects of game theory and reinforcement learning. Moravcik currently serves as CSO and has a Ph.D. in algorithmic game theory, while Kadlec, the CTO, has a Ph.D. in episodic memory modeling and worked on the neural language that is referenced by the GPT family.

Jim Gao and Vedavyas Panneershelvam

Jim Gao and Vedavyas Panneershelvam were among DeepMind’s most senior staff before they resigned to pursue other interests. Gao had previously worked at Google where he led a team that reduced the company’s data center energy costs by 40%, while Panneershelvam had extensive software development experience at Sentenial, Imtech ICT, and Universal Music.

The pair founded Phaidra in February 2020. The company, which in July 2022 announced a $25 million funding round coincidentally led by DeepMind, is a developer of scalable, repeatable, cloud-based AI for industrial purposes. This includes verticals such as refineries, data centers, steel mills, and pharmaceutical plants. 

In an interview with Geek Wire, Gao noted that soaring energy costs had seen demand for the product increase. Phaidra AI can reduce heating or cooling costs by up to 30% and increase profits in the process. “Energy efficiency is no longer nice-to-have,” Gao explained. “It is now business critical.

Jack Kelly

Jack Kelly worked at DeepMind’s London location as a research engineer between June 2017 and December 2018. Specifically, he was part of a team that trained an ML model on weather forecast and turbine data to predict wind power for electricity generation.

Kelly then worked for the International Centre for AI, Energy & Climate in London before launching the start-up Open Climate Fix. The non-profit research lab has a core focus on reducing GHG emissions at scale with open-source solutions.

Kelly was initially pessimistic about using AI to address climate change. But during his tenure at DeepMind, he was inspired by team leader Jim Gao’s optimism and enthusiasm. Today, Open Climate Fix has been backed by notable companies such as Google and Nvidia.

Tejas Kulkarni

Kulkarni was a senior research scientist at DeepMind until November 2019. Before his stint there, he completed a Ph.D. in deep learning, reinforcement learning, and probabilistic programming at MIT. He also briefly interned for Google and worked for Qualcomm in neural network engineering.

Kulkarni later founded Common Sense Machines (CSM), a provider of interfaces, APIs, and open-source software to train AI models to create 3D simulations of the real world. While the scientist described his time at DeepMind as one of the most creative of his life, he relished the chance to move beyond academia and start a company that could make a scalable impact in life.

Darrell Adjei

Darrell Adjei is a software engineer who spent just a year with DeepMind in 2018 and 2019. After a subsequent two-and-a-half-year period at Google as a software developer, payments company Fuse recruited him to work as its founding engineer.

Fuse is a unified API for financial data aggregators and has secured funding from 27 firms including Y Combinator, Long Journey Ventures, and Goat Capital.

Key takeaways

  • According to Business Insider, the DeepMind mafia are a team of 18 alumni from Google’s AI research lab who have raised millions of dollars for various start-ups and non-profits. 
  • Trevor Beck worked in various product management roles at DeepMind over nine years. Beck is now the founder and current CEO of Shift Lab, a company that offers AI marketing and eCommerce solutions to sellers and creators.
  • Former DeepMind colleagues Martin Schmid, Matej Moravcik, and Rudolf Kadlec founded a company that applies AI to the trading industry with aspects of game theory and reinforcement learning. Jim Gao and Vedavyas Panneershelvam were among DeepMind’s most senior staff before they resigned to pursue start Phaidra – a developer of scalable, repeatable, cloud-based AI for industrial purposes.

Key Highlights

  • Karl Moritz Hermann:
    • Former senior researcher at DeepMind’s Berlin office.
    • Co-founded Saiga, an AI-powered personal assistant startup.
  • Trevor Beck:
    • Worked in various product management roles at DeepMind.
    • Founder and CEO of Shift Lab, which offers impactful AI solutions, including Genome, an eCommerce and marketing solution using generative AI.
  • Mustafa Suleyman:
    • Co-founded DeepMind Technologies.
    • Co-founder of Inflection AI, aiming to develop AI software products for human-computer communication.
    • Venture partner at VC firm Greylock Partners.
  • Andrew Eland:
    • Responsible for DeepMind Health’s engineering.
    • Founded Diagonal, a city planning consultancy firm focused on urban data analysis software.
  • Adji Bousso Dieng:
    • Research scientist at Google Brain, focusing on deep learning and probabilistic models.
    • Founder of The Africa I Know, a non-profit supporting African students with educational resources.
  • Peter van Toth:
    • Former AI researcher at DeepMind.
    • Founded Kosen Labs, focusing on AI for web3 and decentralized exchanges.
  • Tiago Ramalho:
    • Former Research Engineering Manager at DeepMind.
    • Founded Recursive AI, a startup building AI models to enhance efficiency and sustainability management.
  • Hugo Penedones:
    • Machine learning researcher and engineer at DeepMind.
    • Co-founded Inductiva Research Labs, applying machine learning to computational problems.
  • Martin Schmid, Matej Moravcik, Rudolf Kadlec:
    • Founders of EquiLibre Technologies, applying AI and game theory to trading.
  • Jim Gao, Vedavyas Panneershelvam:
    • Former senior staff at DeepMind.
    • Co-founded Phaidra, developing AI for industrial energy efficiency.
  • Jack Kelly:
    • Research engineer at DeepMind, focused on climate change solutions.
    • Founded Open Climate Fix, a non-profit research lab addressing climate change with AI solutions.
  • Tejas Kulkarni:
    • Senior research scientist at DeepMind.
    • Founded Common Sense Machines (CSM), providing AI training for 3D simulations.
  • Darrell Adjei:
    • Former software engineer at DeepMind.
    • Joined Fuse, a unified API for financial data aggregators.

Read Next: History of OpenAI, AI Business Models, AI Economy.

Connected Business Model Analyses

AI Paradigm

current-AI-paradigm

Pre-Training

pre-training

Large Language Models

large-language-models-llms
Large language models (LLMs) are AI tools that can read, summarize, and translate text. This enables them to predict words and craft sentences that reflect how humans write and speak.

Generative Models

generative-models

Prompt Engineering

prompt-engineering
Prompt engineering is a natural language processing (NLP) concept that involves discovering inputs that yield desirable or useful results. Like most processes, the quality of the inputs determines the quality of the outputs in prompt engineering. Designing effective prompts increases the likelihood that the model will return a response that is both favorable and contextual. Developed by OpenAI, the CLIP (Contrastive Language-Image Pre-training) model is an example of a model that utilizes prompts to classify images and captions from over 400 million image-caption pairs.

OpenAI Organizational Structure

openai-organizational-structure
OpenAI is an artificial intelligence research laboratory that transitioned into a for-profit organization in 2019. The corporate structure is organized around two entities: OpenAI, Inc., which is a single-member Delaware LLC controlled by OpenAI non-profit, And OpenAI LP, which is a capped, for-profit organization. The OpenAI LP is governed by the board of OpenAI, Inc (the foundation), which acts as a General Partner. At the same time, Limited Partners comprise employees of the LP, some of the board members, and other investors like Reid Hoffman’s charitable foundation, Khosla Ventures, and Microsoft, the leading investor in the LP.

OpenAI Business Model

how-does-openai-make-money
OpenAI has built the foundational layer of the AI industry. With large generative models like GPT-3 and DALL-E, OpenAI offers API access to businesses that want to develop applications on top of its foundational models while being able to plug these models into their products and customize these models with proprietary data and additional AI features. On the other hand, OpenAI also released ChatGPT, developing around a freemium model. Microsoft also commercializes opener products through its commercial partnership.

OpenAI/Microsoft

openai-microsoft
OpenAI and Microsoft partnered up from a commercial standpoint. The history of the partnership started in 2016 and consolidated in 2019, with Microsoft investing a billion dollars into the partnership. It’s now taking a leap forward, with Microsoft in talks to put $10 billion into this partnership. Microsoft, through OpenAI, is developing its Azure AI Supercomputer while enhancing its Azure Enterprise Platform and integrating OpenAI’s models into its business and consumer products (GitHub, Office, Bing).

Stability AI Business Model

how-does-stability-ai-make-money
Stability AI is the entity behind Stable Diffusion. Stability makes money from our AI products and from providing AI consulting services to businesses. Stability AI monetizes Stable Diffusion via DreamStudio’s APIs. While it also releases it open-source for anyone to download and use. Stability AI also makes money via enterprise services, where its core development team offers the chance to enterprise customers to service, scale, and customize Stable Diffusion or other large generative models to their needs.

Stability AI Ecosystem

stability-ai-ecosystem
Scroll to Top

Discover more from FourWeekMBA

Subscribe now to keep reading and get access to the full archive.

Continue reading

FourWeekMBA