Derived from job-description analysis by Serendipath's career intelligence engine.
Original posting from Parallel Bio via Ashby
THE MISSION
More than 95% of drugs that pass animal testing fail in human trials. A single new drug takes about $2.6 billion and 12.5 years to develop, and most of that effort goes to candidates that were never going to work in people.
AI was supposed to change this. But most AI for drug discovery is still trained on the same inputs that got us here: animal data, irreproducible papers, and cell lines that don’t respond like living tissue. A model can only learn the biology it’s shown, and no amount of scale or compute fixes bad data.
What’s missing is data on how living human tissue responds to a drug, across the range of people who will actually take it. That data barely exists, because there has been no way to generate it at scale without putting patients at risk.
Parallel Bio grows it. Our expert biologists build living human immune systems, called immune organoids, from a proprietary biobank spanning ages, sexes, ancestries and disease states. That means we can ask how a drug behaves in an older patient, a specific ancestry or an autoimmune background, not just in an average donor. We've already modeled immunotoxicology, immunogen stimulation and two autoimmune diseases, with more on the way. Our founders are world experts in scalable organoids. Robert is a Stanford-trained immunologist. Juliana trained in the lab that invented the first organoid and later automated them at System1 Biosciences.
We run these experiments in an AI-driven robotic lab, and every experiment makes our world model smarter, which in turn chooses better experiments. We build every layer ourselves: the biology, the robotics, the data infrastructure, and the model that orchestrates it all. The whole lab is callable over API and MCP, so AI agents can design and run real biological experiments, within safety limits we set. For biologists, that means asking questions at a scale and pace no bench can match. For engineers, it means building AI that learns from real human biology instead of the literature.
It's already working. Our platform is 87% concordant with clinical outcomes, compared with roughly 3% translational success for animal models. Five Fortune 500 companies use it to test drugs, and our data has informed a Phase 1 trial. We've raised ~$30M from AIX Ventures, Marc Benioff, Jeff Dean and Y Combinator. With the FDA Modernization Act 3.0 and the FDA's March 2026 validation framework, regulation is moving our way.
The opportunity is generational. We're building the first scaled engine for real human biological data, and using it to know which drugs will work in people before they ever reach a trial.
YOUR SKILLS
Must be an expert in training AI models for biology
Parallel Bio is an equal opportunity employer committed to fostering an inclusive and respectful workplace. We encourage applications from individuals of all backgrounds, regardless of age, gender, ethnicity, religion, disability, or sexual orientation.
Source: Parallel Bio careers (Ashby)