Matthew Jewkes

Matthew Jewkes co-founded MotionHall with Rachael Craig to help biotech companies find licensing partners for individual drugs. At Clarivate, he helped build a system that handled tens of thousands of indexing tasks each month as drug-development records were added or changed. One of his team’s most consequential decisions was to let it leave a question open. Sometimes an apparent error meant that two experts agreed on the biology but disagreed about how to index it. Matthew later co-founded Standard Cybernetics.

From a Halifax software cooperative to biotech licensing

In late 2009, Matthew and Devin Horsman founded Twisted Oak Studios in Halifax. Game studios brought the cooperative engineering problems, and the team developed particular expertise in virtual reality and the Unity game engine. Matthew also contributed to the first iPhone release of Signal as part of the Twisted Oak team. Released in 2014, the app let people use their existing phone numbers to make encrypted calls, including to RedPhone users on Android. Open Whisper Systems later invited Matthew to help grow its developer community.

Matthew met Rachael at Twisted Oak. When River Studios bought the studio in 2015, Devin stayed to lead the Canadian operation. Matthew negotiated terms that let him start a company with Rachael. Their early work together produced a tool for tracking scientific research and the products emerging from it. Twisted Oak’s sale · MotionHall’s origins

At MotionHall, Rachael served as chief executive and Matthew as chief technology officer. Their OutMatch software used data about drug candidates and prospective partners to help dealmakers decide which companies to approach for a licensing agreement. In March 2024, Clarivate announced an agreement to acquire most of MotionHall’s assets and said the team had joined its life sciences business.

By July 2025, Matthew was Clarivate’s senior vice president for AI product transformation in life sciences and healthcare. He argued that a licensing decision requires knowledge that even a detailed database of drugs and deals can miss. A therapy might look promising on the clinical evidence, but could the buyer’s manufacturing partner produce it through Phase 2 trials? He wanted decision-support software to bring that internal knowledge together with scientific, regulatory and market evidence while a team was weighing a deal. Matthew’s essay on biotech decision-making

Which indexing decisions need an expert?

At Clarivate, Matthew and his colleagues first tried an agent that could research a record and assign it to the index. They brought about 1,000 evaluation examples to domain experts. One question illustrated the limits of an accuracy score. Should a drug’s targets include a molecule it affects indirectly, or only the protein it binds to? Correcting a model error required a known answer. Resolving this question required an indexing policy.

The team narrowed the agent’s role. Routine assignments could run without human approval. The system set aside disputed classifications, referred unfamiliar science to experts and sent business-critical assignments through a separately tested workflow. Matthew reported 99.95% accuracy on the business-critical tasks after separating out ambiguous and unusual cases. The team deployed the system only after a random sample showed that it outperformed the historical indexing work. Matthew’s account of the Clarivate system

Matthew favors that investment for established processes where the volume of work and the cost of mistakes justify the engineering effort. For a team still discovering what customers need, he favors keeping the work flexible. At Clarivate, grouping the exceptions gave experts a way to do both. They could recommend shelving a recurring dispute that mattered little to customers, or spot demand in cases they had previously seen only weeks apart. One expert had already worked out where to invest by the time a major customer raised the same need.