Company

Dr Elise Park leads BestSoft AI's work on model safety, alignment and responsible deployment. She oversees the teams responsible for evaluating how advanced systems behave under uncertainty, how reliably they follow human intent, and how emerging capabilities should be tested before they move into wider use.

Elise is known inside BestSoft for asking the question that tends to arrive immediately after everyone else has become excited about a result: what happens when this goes wrong? Her approach is methodical rather than alarmist. She believes capable AI systems can be enormously useful, but that usefulness depends on understanding their limits as carefully as their strengths.

She works closely with research, engineering and policy teams to ensure safety considerations are incorporated throughout development rather than added at the end.

Dr Arjun Iyer leads BestSoft AI's core research organisation, overseeing work on frontier models, reasoning, agentic systems and new model capabilities.

A British-Indian researcher with a reputation for following interesting results considerably further than originally planned, Arjun represents the capability-driven side of BestSoft's research culture. He is deeply interested in what advanced systems can become, while accepting that every increase in capability creates new questions for the safety teams sitting a few doors away.

That makes his working relationship with Elise Park particularly important. They frequently approach the same result from opposite directions: Arjun wants to know what else the system can do; Elise wants to know what else it might do.

He is analytical, curious and unusually comfortable saying that a result has surprised him.

Clara Nunes leads BestSoft AI's global affairs, public policy and government engagement work.

Her team works with policymakers, regulators, researchers and civil-society organisations on the governance questions emerging around increasingly capable AI systems. Clara's role is not simply to explain BestSoft's research externally, but to translate between people who often approach the technology with very different assumptions, incentives and vocabularies.

She is precise, diplomatic and exceptionally difficult to push into saying more than she intended. Internally, she is valued for identifying the second- and third-order consequences of decisions long before they become public-policy problems.

Clara works closely with both the research and safety organisations, particularly where technical capabilities begin to raise questions that cannot be answered by engineering alone.

James Chen leads the engineering, infrastructure and systems teams responsible for turning BestSoft AI's research into reliable technology.

His remit covers model infrastructure, platform architecture, large-scale compute systems, deployment reliability and the technical systems that allow research models to operate safely outside the laboratory. Where the research organisation asks whether something can be built, James's teams are responsible for determining whether it can be built reliably enough to depend on.

He has a pragmatic engineering philosophy: sophisticated systems are useful only when they remain understandable, maintainable and predictable under pressure. He is particularly focused on reliability in agentic systems, where failures can emerge from interactions between models, tools, infrastructure and humans rather than from any single component.

James is calm, analytical and notably resistant to technological theatre. He tends to reduce complicated arguments to the engineering problem underneath them.