Distilling AI solutions for enterprise
We enable enterprises to build smaller, more focused AI models with greater performance at a fraction of the cost.
The AI era has just started
Unlock the value AI has to offer
Existing Large Language Models (LLMs) are overly general and lack specificity. Malted AI works with enterprises to develop highly specialised solutions that solve the hardest domain-specific problems.
How we work
Our partnership model
We combine enterprises’ domain expertise with our world-class knowledge distillation technology to solve the highest-value problems. We distil enterprise-specific Small Language Models (SLMs) to achieve pinpoint accuracy with 10 – 100x cost savings.
Our technology
Knowledge Distillation
High-quality data from a “teacher” system is used to train a network of proprietary “student” SLMs that are optimised for a single problem.
Opposed to general AI that does thousands of tasks moderately well, Malted AI’s SLMs do one task near perfectly.
Use Cases
Search and understanding
Find answers to domain-specific questions over company knowledge.
Report
generation
Support decision making through automated reports and insights.
Information extraction
Real-time extraction of structured information from customer interactions.
Content generation
Personalise the way you communicate with your customers.
With us, you can expect
Higher
Performance
Achieve ROI on high-value problems where general AI fails.
Custom Design
A partnership model that builds tailored solutions to enterprise problems
Cost-effective
Smaller, more efficient models that reduce costs by a factor of 10-100x.
Security
Deploy safely on clients’ VPCs, on-prem, or the Malted AI secure cloud.
Meet Our Founders
Our founding team are leading PhDs that previously won the multi-million-dollar Amazon Alexa Challenge, beating more than 100 teams around the globe. Since then, they have raised investment from leading VCs to build one of Europe’s leading AI company.
Iain Mackie
CEO
Carlos Gemmell
CTO
Federico Rossetto
Chief of Engineering
View our career opportunities
Software Developer
Senior Front-End Developer
Senior UX/UI Designer
Head of Business Development
Drive the development of our platform and support the deployment of Machine Learning models at scale.
Shape the face of our AI platform. We’re looking for a creative individual passionate about building intuitive and dynamic user interfaces.
Support the development and usability of our product D-RAG crafting intuitive, aesthetical and functional digital experiences for our users
Play a key role in driving our growth strategy and expanding our market presence.
Software Developer
Drive the development of our platform and support the deployment of Machine Learning models at scale.
Senior Front-End Developer
Shape the face of our AI platform. We’re looking for a creative individual passionate about building intuitive and dynamic user interfaces.
Senior UX/UI Designer
Support the development and usability of our product D-RAG crafting intuitive, aesthetical and functional digital experiences for our users
Head of Business Development
Play a key role in driving our growth strategy and expanding our market presence.
Frequently Asked Questions
Small Language Model (SLMs) vs Large Language Models (LLMs)?
Both LLM’s and SLMs are advanced AI systems that have undergone extensive training using massive volumes of text data, making them capable of understanding existing content and producing new, original content. However, there are differences between these models in relation to their size, performance and efficiency.
LLMs as the name indicates it, are bigger, which means more computational requirements and processing power is needed. Making SLMs faster and cheaper to run in comparison.
LLMs are trained using diverse data and optimised for multiple tasks, in contrast, SLMs are optimised for single tasks using domain-specific data.
While LLMs may perform better across multiple general scenarios, the use of SLMs will increase the performance and efficiency for domain-specific tasks, e.g., monitoring regulatory compliance, or answering complex questions from patents.
What is knowledge distillation?
Knowledge distillation is a process in machine learning where the knowledge from a larger, more complex model (teacher) is transferred or distilled into a smaller, more efficient model (student).
At Malted AI, we are experts at optimising a network of student SLMs based on high-quality domain-specific data produced by the teacher system. This enables our solutions to automatically adapt to specific enterprise problems and continuously improve.
Get in touch
Book a call to discuss how Malted AI can help.
Address
33 Melville Street
EH3 7JF
Edinburgh
Scotland 🏴