On Thursday, TitanML, a London-based AI startup, revealed that it had raised $2.8 million (or around €2.6 million) in a pre-seed round of fundraising from deep tech-focused angel investors and Octopus Ventures.
The financing comes as Titan Takeoff, a UK startup that dramatically speeds up and reduces the cost of large language model (LLM) deployment for machine learning teams, launches.
“We see a ton of organizations making significant investments in AI, but the majority have trouble seeing a return on their capital.
Machine learning teams deal with various issues when deploying models, from GPU shortages to privacy concerns when using third parties like OpenAI. According to Meryem Arik, CEO and co-founder of TitanML, “We’re changing what was formerly the most difficult phase of the development cycle into the quickest and most efficient, and it’s here that our flagship product, Takeoff, comes into its own. The company aims to provide the foundation for deep learning in businesses.
TitanML: Important information
TitanML was established in 2021 due to the postdoctoral study of Drs. James Dborin and Fergus Finn at University College London (UCL) investigated deep learning training, compression, and inference optimization.
Later, they teamed up with Oxford alumna and co-founder Meryem Arik.
TitanML enables machine learning teams to deploy large language models (LLMs) quickly and effectively. Expert ML researchers, engineers, and architects make up the team now trying to hasten the implementation of deep learning in businesses.
The Financiers
Octopus is a multi-stage venture finance firm focusing on B2B software, biotechnology, the environment, consumer products, deep technology, and health. The VC invests more than £200 million annually to support the rapid scaling of its portfolio firms.
According to Mat Munro of Octopus Ventures’ Deep Tech Team, “new language models are currently the subject of much interest. However, the truth is that for businesses looking to implement them internally, it may require months of labor from highly qualified data scientists who still need to optimize these models to make them usable products.
TitanML lets businesses quickly, cheaply, and with as few trade-offs as possible, greatly reduce the processing needs of high-specification models, which lets these technologies reach their full potential. He continues, “We are thrilled to be leading this round of funding and look forward to helping the team as they take the next stages in enabling businesses with effective model implementation.
