Open role

Co-founder & CTO, AI for predictable drug discovery and development

Elman is building AI to predict clinical trial and drug development outcomes, and quantify how much confidence those predictions deserve. Our starting point is the evidence available when a decision is made, what we judge from it, and what happens next. By linking those records, we aim to learn which experimental findings predict outcomes in which therapeutic settings. The engine already gathers evidence and tests scientific explanations. We're looking for a technical co-founder to lead its development and work out how to measure and improve its predictive performance.

London or remote, flexible Full time Pre-seed, spinning out of Deep Science Ventures Founder equity

Apply →

What we are working on

How reliably different kinds of biological evidence predict human outcomes, and whether that can be measured at all.

Drug discovery and development involve decisions about targets, treatment approaches, delivery and patient populations. Each decision depends on evidence that may not transfer well to the setting that matters: treating people. A treatment working in mice, for example, does not tell us how much confidence to place in its chances in humans.

Around nine in ten drugs entering clinical trials never reach approval. The causes run from efficacy to safety to commercial choice, and the claim here is narrower than blaming that number on translation: the field has never systematically measured how far any given kind of evidence should carry.

We want to measure that. It is difficult to test. Outcomes can take years to arrive, suitable labels are often missing, and a failed trial does not necessarily reveal which earlier assumption was wrong. There is no established dataset or method that settles the question, so defining what to measure, building reliable evaluations and choosing appropriate methods are central to the role.

Read the full thesis →

Approval likelihood from Phase 1 is roughly 10%. Hay M, Thomas DW, Craighead JL, Economides C, Rosenthal J, Clinical development success rates for investigational drugs, Nature Biotechnology 32(1):40-51, 2014.

Your role

The founding team covers biology and commercial development. You will be responsible for the technology and the research needed to evaluate it. There is no engineering team yet: you will be building the system yourself and should expect to remain hands-on for at least the first year.

Developing the platform. Take ownership of the existing system, set technical priorities and implement improvements.
Building the measurement programme. Define what to predict and how to assess it, develop suitable datasets and methods, and report uncertainty alongside results. Identify where the system is reliable and where it is not.
Preventing misleading evaluations. Address leakage, label quality and contamination. Retrospective tests need particular care when an LLM may have encountered the outcome during training.
Hiring the first technical team once funding is in place: computational biologists to widen data acquisition, and engineers to run evaluation.
Supporting fundraising. Explain the technical approach, evidence, limitations and development plan to investors.

The modelling approach is open. You will choose methods based on the problem and the available data, whether those involve Bayesian or hierarchical inference, graph learning, calibration, classical statistics or a combination.

How we think about contamination →

What we are looking for

Core

Evaluation and benchmark design. Experience defining labels where they are not readily available, checking their reliability and designing data splits that prevent leakage.
Applied statistics. A strong understanding of calibration, confidence intervals, base rates and the limits of small or noisy datasets.
Breadth across machine-learning approaches. Hierarchical and Bayesian inference, classical statistics, deep learning on graphs including graph neural networks, probabilistic graphical models and learned calibration. We care less about depth in any one of these than about how you decide between them on a given dataset.
Production LLM systems. Agent orchestration, tool integrations, context management and evaluation systems that catch regressions. The platform is model-agnostic and routes across Anthropic, OpenAI, Google and specialist models, so experience running more than one provider in production is useful.

How you work

Hands-on technical leadership. You lead development and build the system yourself. AI coding assistants and automated pipelines run through most of our work. You would own how we use them, and be responsible for reviewing and testing what they produce.
Careful interpretation of results. You report uncertainty and limitations, and recognise when the available evidence is insufficient to support a conclusion.
Clear technical communication. You can explain and defend your decisions to collaborators and technical investors.
You write production Python.
Experience with biological or clinical data would be useful, but a biology background is not required. Previous founding experience, and research, software or open-source work we can review, are also welcome.
Location is flexible. The company is UK-based and we prefer candidates here or willing to relocate, but we already work distributed and remote is workable for the right person.

Stage

Elman is a pre-seed company, and the spin-out from Deep Science Ventures completes over the coming months. The platform is already running and has produced a scientific result through work with the Allen Institute, described in a public preprint. The next funding round will fund the measurement programme.

Worth reading before you apply: the thesis, the preprint, and the engine running.

Questions people ask first

Do I need a biology background?

No. The founding team covers biology. Experience with biological or clinical data is useful, and what matters more is the instinct to ask whether a measurement means what it claims.

Is this remote?

Location is flexible. The company is UK-based and we prefer candidates here or willing to relocate, but we already work distributed and remote is workable for the right person.

What stage is the company at?

Pre-seed. Elman was built inside Deep Science Ventures and is spinning out. The platform is already running and has produced a scientific result through work with the Allen Institute, described in a public preprint. No round has closed yet.

How do you work with AI coding assistants?

They run through most of our development, along with automated pipelines around them. You would lead that work and own how we use and improve these tools.

Would this be called Head of AI somewhere else?

Possibly. The work is what a Head of AI or founding engineer does at a larger company, plus the research programme and the fundraising. Here it is a co-founder seat with founder equity, which is the material difference.

Is this a CTO role or a co-founder role?

Both. You would join as a co-founder, with founder equity and technical authority from day one. The title is CTO because that is what the job is.