25 FKJVkZ4 vTsBTp39Uz - Radiology — California, United States

mz rh6MHgS H6RHsppT28 – Radiology (In Partnership with LatchBio)

Full-Time | Remote | United States

We’re partnering with LatchBio, a rapidly growing AI-driven biotechnology company, to recruit an AI MyYwPg6 E8AUt473GD specialising in Radiology.

LatchBio builds rigorous scientific benchmarks that help leading AI labs and pharmaceutical companies evaluate how effectively their AI agents can perform real-world biological and clinical research workflows.

Their Radiology team is building RadiologyBench, which evaluates AI capabilities across diagnostic imaging and image-based clinical decision-making, spanning MRI, CT, ultrasound and radiography.

AI agents can now work through an imaging study the way a radiologist does: open the dataset, explore it based on an indication, identify relevant features and validate findings.

This is an opportunity to apply your hands-on imaging expertise at the intersection of radiology and frontier AI, helping establish the standards against which advanced AI systems are evaluated.

The Opportunity

You’ll decide what the right answer is for each case, using your hands-on experience reading or acquiring medical images to design challenging evaluations and assess whether AI-generated conclusions are genuinely supported by the imaging evidence.

Working alongside scientists and engineers, you’ll translate real radiology workflows into structured tasks and benchmarks, from whether a scan was acquired properly to what it shows. You’ll then review how AI agents actually handled each case, including what they missed and what they reported that isn’t there.

This is a clinical-first position where deep imaging experience, sound judgement and the ability to challenge flawed conclusions are more important than advanced programming expertise.

Key Responsibilities

• Write the questions radiologists frequently ask during interpretation of medical images, across MRI, CT, ultrasound and radiography and across organ systems.

• Decide the optimal answer format the agent should use and what counts as a pass given the objective.

• Establish the correct answer for each case, accounting for inter-reader variability, imaging uncertainty and legitimate expert disagreement.

• Review what agents are actually doing and whether their approach holds up, including what they missed and what they reported that is not present in the images.

• Distinguish true findings from image problems, including artefact, positioning errors and incorrect acquisition parameters.

• Evaluate whether AI agents can tell the difference between technical correctness and clinical soundness, and between a confident read and a well-calibrated one.

• Help decide what the agent actually needs to solve each objective, and work closely with scientists and engineers to improve the benchmarking workflows.

What We’re Looking For

• Professional experience working with medical images, with thousands of studies handled. Your role could include:

• Radiologist, including those trained outside the US (board-certified or equivalent)

• Radiology resident or fellow

• Sonographer (ARDMS or equivalent)

• MRI or CT technologist (ARRT or equivalent)

• Deep experience in at least one modality (MRI, CT, ultrasound or radiography) and working familiarity with the others.

• Knowledge of normal anatomy and common pathology for your area, and the ability to tell when the image itself is wrong: artefact, positioning or parameters.

• Enough comfort with Python to work on interpretation and run scripts. ML engineering experience is not required but could come in handy.

• Daily use of AI tools such as Claude Code or similar agents, with the judgement to distinguish between technical correctness and clinical unsoundness.

• Confidence in challenging conclusions, explaining diagnostic uncertainty and updating your position when presented with new evidence.

• Bonus: imaging AI challenge experience (e.g., RSNA, MICCAI) or contributions to resources such as Radiopaedia.

• Bonus: familiarity with public imaging archives (TCIA, MIMIC-CXR, UK Biobank, OpenNeuro).

Why Join LatchBio?

• Frontier AI exposure: Help shape how advanced AI systems are evaluated across medical imaging and clinical research.

• Meaningful scientific impact: Apply your expertise to challenges involving image acquisition, interpretation and diagnosis.

• Scientific ownership: Become the authority setting the standards for how AI agents are evaluated against real-world imaging practice.

• High-growth environment: Work with a company collaborating with leading AI laboratories and pharmaceutical organisations.

• Remote flexibility: Join an internationally distributed team, with LatchBio headquartered in San Francisco.

• Interdisciplinary collaboration: Work alongside scientists and engineers at the intersection of medicine and artificial intelligence.

Location & Working Arrangements

Location: Fully Remote | United States

Employment: Full-Time

LatchBio is headquartered in San Francisco, with the Radiology team being built remotely across international markets.