Advanced Computing in the Age of AI | Thursday, February 22, 2024

TitanML Secures $2.8M to Solve the LLM Deployment Nightmare Plaguing Machine Learning Teams 

LONDON, Oct. 12, 2023 -- TitanML has secured $2.8 million in pre-seed funding led by Octopus Ventures alongside deep tech-focused angel investors. The round is timed as the team publicly launches Titan Takeoff, which makes large language model (LLM) deployment significantly faster, cheaper and easier for machine learning teams.

Founded in 2021, TitanML was born out of the postdoctoral research of both Dr. James Dborin and Dr. Fergus Finn, who explored deep learning training, compression, and inference optimization whilst at University College London (UCL). Together with Oxford alumnus, Meryem Arik, the trio transformed their expertise into enterprise-ready software which slashes the effort of both AI development and deployment.

“We see so many businesses investing heavily in AI, but most struggle to return value from their investments. Machine learning teams are facing a whole host of problems in model deployment - from GPU shortages, to concerns over privacy when deploying through third parties like OpenAI. We’re transforming what was once the most challenging part of the development cycle into the easiest and most efficient, and it’s here where our flagship product, Takeoff, comes into its own,” said Meryem Arik, CEO & co-founder of TitanML.

To date, the team’s achievements include the real-time deployment of state-of-the-art Falcon LLM on a commodity CPU - a feat which garnered significant industry recognition. TitanML has also secured key strategic partners including Intel and AWS, alongside a series of enterprise clients after demonstrating up to 90% reductions in compute costs and 20x latency improvements within just hours of deployment.

Commenting on their investment in TitanML, Mat Munro of the Deep Tech Team at Octopus Ventures said: “There’s a lot of excitement at the moment around new language models. For companies wanting to deploy these in-house however, the reality is that it can take months of work by highly skilled data scientists that still need to optimise these models to turn them into deployable products. TitanML makes it possible for businesses to significantly reduce the processing requirements of high specification models quickly, cheaply, and with minimal trade-offs, unlocking the true potential of these technologies.”

“We are very excited to be leading this round of investment and look forward to supporting the team as they take the next steps in empowering businesses with efficient model deployment,” he added.

As the adoption of AI and LLMs grows exponentially, TitanML’s ambition is to build the infrastructure layer for deep learning in the enterprise. As the company creates a platform designed to become the gold standard for LLM deployment, it is fast delivering on a promise to help machine learning engineering teams realise the value of their AI investments.

About TitanML

TitanML enables machine learning teams to effortlessly and efficiently deploy large language models (LLMs). Founded by Dr. James Dborin, Dr. Fergus Finn and Meryem Arik, and backed by key industry partners including AWS and Intel, TitanML is a team of dedicated deep learning engineers on a mission to supercharge the adoption of enterprise AI.


Source: TitanML

EnterpriseAI