London-based Bitfount has raised $5 million in a seed round led by Ahren and Speedinvest. The startup addresses the complexities around collaboration between two or more parties when working with highly sensitive data, e.g. moving algorithms to data instead of the other way around. The new capital will be used to further compute product development and expansion of the engineering and design teams.
While the global demand for machine learning/AI and analytics in just about every industry continues its skyward trajectory, running in tandem is increased public and regulatory awareness and scrutiny of data privacy.
Founded by former Chief Architect for Apple’s Siri Understand, Blaise Thomson, and CPTO at DueDil, Naaman Tammuz, Bitfount argues that AI demands and data privacy concerns can travel in pairs, it all depends on just how much, and of what is shared with whom.
“Where traditionally the legal process behind sharing raw data between parties can take months or even years, Bitfount’s secure network instead allows algorithms to travel to datasets rather than the other way around, which means there needn’t be any disclosure of sensitive information,” explained Thomson. “This new paradigm unlocks opportunities for both analysis and machine learning that were previously impossible - either because they required sharing of sensitive raw data, or because the contractual agreements were prohibitively complex.”
Now if this sounds like federated AI and data collaboration, you’d be right. However, Bitfount aims to stand apart with its secret sauce: a novel message-based architecture that according to the company makes it much easier and faster for IT teams to deploy.
The ‘simple and ease of use’ factor is key, backed up by a hybrid open-source/SaaS model, and an SDK backed by a web-based platform that includes a catalogue for dataset discovery, flexible and granular role-based access controls, full audit history, support for custom models, algorithms, and data sources, and cryptographic tools such as differential privacy and secure multi-party computation (SMPC) for additional privacy guarantees.
Bitfount says that while it’s seen interest from a number of industries, right now the team is catering to healthcare, an area where privacy concerns and data collaboration hurdles are paramount. Thomson specifically pointed to areas including patient identification for clinical trials, digital biomarker development, drug development, AI/ML model evaluation for regulatory and benchmarking purposes, and improving data diversity and representation for better healthcare outcomes.
While the company remains tight-lipped about its existing client base, it does point to, “a number of pharmaceutical companies, research institutions, hospitals, and major financial services firms.”
On the investment, Speedinvest’s Rick Hao commented, “Bitfount is strategically placed to create an ecosystem in which dataset owners can safely be analysed by external users. It is democratising the technology to fully unlock the power of data, and we are certain that many different industries will look to adopt their best-in-class product.”