AI Domain Researcher

Role Summary

Unlock AI in safety-critical domains

AI Domain Researchers at HASH explore challenges and use cases applying AI in safety-critical domains. HASH is exclusively focused on developing AI solutions that provide "at least" expert levels of accuracy, and extremely high confidence outputs – including data, predictions and decisions – which are both safe and reliable.

Researchers play a central role in realising the transformational potential of AI, by working with domain experts to develop use cases and deployment pathways for frontier AI technology. This involves specifying models of the environment, the goals and safety rules for each use case, and identifying and solving barriers to successful deployment.

We are currently recruiting researchers across supply chain logistics and life sciences as part of an Advanced Research + Invention Agency (ARIA)-funded project, undertaken in conjunction with Department of Computer Science at the University of Oxford.

Responsibilities

Codify problem spaces

Researchers spend a large amount of time talking to users, and may even be embedded within partner organizations from time-to-time. As representatives of HASH in a given area they possess deep knowledge about that domain, as well as strong personal connections throughout it. Researchers are primarily responsible for codifying relevant problem spaces in a way that is useful for the development of new generative and agentic AI systems which have the potential to significantly impact operations in safety-critical systems.

Typical tasks will include:

Interviewing users to gather requirements and evaluation criteria for potential use cases
Building and maintaining domain-specific ontologies
Working with users to secure access to datasets for evaluating and providing context to AI systems
Gathering and analyzing information to produce new, novel datasets
Developing and iterating on domain-specific models
Constructing evaluations for the safeguarded deployment of AI
Conducting experiments in-silico and deploying models in the real world
Documenting findings in publishable papers and reports
Presenting research outputs to a varied range of stakeholders

Requirements

Deep domain expertise

Researchers are experts in one or more safety-critical domains, and work closely with both real-world practitioners and academics to study these spaces and the potential impacts of AI solutions within them. Researchers must be highly self-motivated, comfortable talking to domain practitioners and seeking new contacts to help further the research, experienced in publishing or presenting original research in a world-class environment, and capable of working under strict time pressures.

Researchers work with potential partner organizations to secure access to the necessary data required to produce domain models, working with them to define a credible path to deploying new AI tooling. Commercial awareness, aptitude for navigating corporate and political barriers, and comfort interfacing with external stakeholders are therefore all required.

A key part of the role will be gathering raw information from domain experts about the challenges they face and translating this into a structured definition of processes, contextual information, and rules within the domain, suitable for translating into more formal definitions to be consumed by an AI model. This will involve knowing what questions to ask in order to be able to fully define the causal relationships between different actors in the system, the data points that might meaningfully contribute to AI decisions or predictions, and how the impact of AI tools can be measured and evaluated – including safety rules and other boundaries that must be respected. Any familiarity with process mapping, world modelling and similar tasks will be highly beneficial.

What You’ll Achieve

Safeguarded AI

You'll attend (and potentially even speak at) conferences, and publish academic research.

You'll be contributing directly to the development and first deployments of new AI technology.

You'll produce novel, comprehensive analyses of key problems within your domain.

Our mission

Overcome bounded rationality

Learn more about our mission →

Our culture

Intense by design

We're focused on building something big, in the service of something important (our mission). To that end, we're unapologetically excited. Actions speak louder than words, and we measure performance by output. You can work from wherever, but we know that teams are more productive together, in-person. We prioritze speed, and measure product delivery timelines in hours and days, not months and years. We value high-energy, high-expectations people who do what they say and say what they mean.

Our benefits

Life at HASH is good

We provide leading equity-weighted total comp, and depending on your location also offer:

Competitive salary and meaningful equity
Tax-advantaged options (e.g. EMI in the UK)
Employer pension contributions
Unlimited holiday
Quarterly in-person team retreats
Gym membership/workout classes
Fully-funded HSA & comprehensive health insurance

Interested?

Let us know you exist

We don’t accept applications to specific roles at HASH, but you can tell us about yourself to make sure you’re on our radar, and share why you're interested in our work. We review every message received, so be sure to include specific examples of relevant experience.