Ghostboard pixel

Subscribe to Our Newsletter

Success! Now Check Your Email

To complete Subscribe, click the confirmation link in your inbox. If it doesn't arrive within 3 minutes, check your spam folder.

Ok, Thanks

Malted AI raised £6M in seed funding for its knowledge distillation platform

Malted AI raised £6 million in a seed funding round led by Hoxton Ventures, and backed by Creator Fund and angel investors. The startup is developing a knowledge distillation platform to create of domain-specific small language models that can accurately and cost-efficiently tackle a single task.

Ellie Ramirez-Camara profile image
by Ellie Ramirez-Camara
Malted AI raised £6M in seed funding for its knowledge distillation platform
Credit: Hoxton Ventures

Malted AI recently announced it has raised a £6 million seed funding round led by Hoxton Ventures, and backed by Creator Fund and angel investors. Founded by a former quant trader, a former Meta researcher, and the former leader of a talented LLM and applied AI team, Malted AI is a company developing a knowledge distillation platform that creates domain-specific small language models (SMLs) that aim to solve a specific problem with unparalleled efficiency and accuracy.

By now, it is a relatively well-known fact that, in addition to their significant costs, large language models (LLMs) are too general by nature, as they were designed to fare acceptably well in a wide variety of tasks. Tailoring options like fine-tuning and retrieval-augmented generation provide strategies to increase the efficiency of LLMs in business use cases. However, those can be costly too, or require technical expertise that may not be available at that point in an organization. To address this challenge, Malted AI proposes a platform that works with enterprises to develop SLM-based solutions that can solve their domain-specific problems.

The SLMs at the core of Malted AI's solutions are trained using a technique known as knowledge distillation, where domain-specific SLM networks are paired with a larger model that can "teach" the network everything it knows on a specific task, so the SLMs in the network can refine their capabilities an perform the relevant task near perfectly. This approach was born of the concepts that earned the founders the 2022 Amazon Alexa Prize, where Iain Mackie, Carlos Gammel, and Federico Rossetto beat over a hundred teams to win one of the largest AI competitions.

According to the company, the technology is already being tested in a pilot phase, and Malted AI plans to continue the technology's rollout through the end of 2024. Hoxton Ventures Partner Bryan Gartner lauded the founding team's talent and connections to the Universities of Edinburgh and Glasgow. He also highlighted the trust that the firm was placing in the startup to positively impact the adoption of AI in enterprise settings. Malted AI will leverage the funds to accelerate most of its operations, as the startup is planning to invest in research, product development, and recruitment as part of its growth strategy.

Ellie Ramirez-Camara profile image
by Ellie Ramirez-Camara
Updated

Data Phoenix Digest

Subscribe to the weekly digest with a summary of the top research papers, articles, news, and our community events, to keep track of trends and grow in the Data & AI world!

Success! Now Check Your Email

To complete Subscribe, click the confirmation link in your inbox. If it doesn’t arrive within 3 minutes, check your spam folder.

Ok, Thanks

Read More