Staff Engineer / Tech Lead

The role

Every new medicine that reaches a patient has passed through a clinical trial. For many of the people in those trials, it is not one option among many. It is the last one. The drug being tested is the thing standing between them and a disease that has run out of other answers.

And almost every one of those trials is still run the way it was decades ago: people reading hundred-page protocols by hand, then combing through site data line by line, hoping to catch the safety signal or the protocol deviation that matters before it becomes the thing that matters too much.

Nearly 90% of trials fail — and around 40% of those failures are not about the science at all. They are operational: the missed deviation, the data caught too late, the oversight that was always going to be too manual to hold.

That gap, between how much is at stake and how manual the oversight still is, is where patients lose their last chance and where good drugs get delayed or buried.

Espresso closes it. We read the protocol so a machine understands it, turn it into the risk indicators a trial should be watched against, and surface the deviations and safety signals that need a human, while filtering out the noise that doesn't. Clinical oversight that is risk-based in practice, not just on paper.

This is the other half of why we exist. AI has arrived as a genuinely new way of working — LLMs and agentic workflows that can read, reason, and act. Almost every industry is being rewritten by it. Clinical trials, where the cost of being wrong is highest and the tooling is oldest, are exactly where it is needed most and slowest to arrive. Espresso is bringing that era to clinical trials, carefully, on real data, in a domain where being right is regulated, audited, and non-negotiable.

This is not a dashboard company. We are building the intelligence layer for how trials are watched.

We Are Early

• Three design-partner sponsors
• A small and senior team
• Real data flowing
  
  

The person who joins now decides what Espresso becomes. We are looking for a Staff Engineer / Tech Lead who can hold both the weight of the mission — lifesaving medicine reaching the people who have run out of time — and the craft of the moment, where the product is still being defined feature by feature.

What you'll do in year one

• Set the technical direction. Own the architecture of the systems you lead. Define the patterns, draw the boundaries between services, and decide what is worth building versus buying. Make the call clearly, write down the reasoning, and revisit it when reality disagrees. Hold the line on what is durable versus what is a shortcut that turns into debt, and know when the shortcut is the right call anyway.
• Build the hard parts and know what not to build. You are hands-on. The thorny multi-tenant isolation, the AI extraction and deviation pipeline, the data model that has to stay correct as it grows — these are yours to design and ship. You set the standard by writing code people want to read.
• Make the AI trustworthy. We run LLMs over clinical protocols and trial data. You will help build the layers that make model output safe to act on: validation, evaluation, guardrails, observability, and the honest assessment of what the model can and cannot be trusted to do. Impressive is easy. Reliability is the job.
• Build correctness in. In this domain, correctness, auditability, and security are not separate workstreams. They are properties of well-built software. You will help make sure the platform stays right under load and under audit: tenant isolation that holds, an audit trail that answers "who changed what, when," encryption and access controls that are real and not aspirational.
• Lead and level up the team. Staff is a force multiplier, not a solo act. You will review code with care, mentor without condescension, give clear context and honest feedback, and make the engineers around you better. You set the bar by example, and you defend it.
  
  

What we're looking for

• 8+ years building production software, with deep ownership of complex systems end-to-end.
• Strong with our stack or its close cousins: TypeScript, Node/NestJS, React, PostgreSQL, and cloud-native infrastructure (AWS, Kubernetes, Terraform). You ramp up fast.
• You have built and led at an early stage. You make good decisions when requirements are moving, and the data is messy.
• You know how to take tasks end-to-end, even when not everything is clear and already decided. You know how to work in uncertainty.
• You think in systems: states, edge cases, failure modes, durable patterns versus one-time fixes that become debt.
• You have shipped AI/LLM-backed features and understand what models can and cannot do, and how to engineer around that honestly.
• You care about correctness and security as a craft, not a checklist, especially when real, sensitive data is on the line.
• You are direct. You give feedback clearly, take it without defensiveness, and say "I don't know" without embarrassment.
  
  

Bonus points

• Experience in a regulated or safety-critical domain (healthcare, clinical, fintech, infrastructure) where being wrong has real consequences.
• Multi-tenant SaaS architecture and data-isolation experience.
• Hands-on work with LLM pipelines: evaluation, output validation, prompt and retrieval design, cost and latency tradeoffs.
• Comfort owning infrastructure and CI/CD, not just application code.
  
  

What we offer

Espresso is building the AI risk-monitoring layer for clinical trials. We are a small, senior, motivated team — early enough that the systems you build now become the foundation, and serious enough that real sponsors trust us with real trial data. We move fast and hold a high bar, because in this domain the bar is the product.

A note on fit: We are not looking for someone who has done this exact thing before. We are looking for an engineer built for this kind of problem — one that sits at the intersection of AI, real data, and an industry that has every reason to be careful. If building software where correctness genuinely matters sounds like the right problem for this moment in your career, reach out.