AI Biologist - Radiology
About The Position
Latch builds scientific benchmarks for AI agents across biology and is developing RadiologyBench to evaluate how well AI agents handle real medical-imaging workflows. The role is for professionals who read or acquire medical images and who can judge image quality and clinical findings with expert care. You will determine the correct answer for each case and build studies that assess agents, from whether a scan was acquired properly to what it shows. You will review agent performance, including missed findings and reported findings that are not present. The resulting evaluations help research labs understand where their models fall short of expert performance. Work spans MRI, CT, ultrasound, and radiography across organ systems. You will design interpretation tasks by writing questions commonly asked by radiologists, choosing suitable answer formats, defining pass criteria, reviewing agents’ approaches, and identifying what agents need to accomplish the objective. The role requires professional experience with medical images, strong expertise in at least one modality and working familiarity with others, and the ability to use Python to interpret results and run scripts. No specific ML engineering experience is required. Latch describes its culture as transparent, collaborative, and respectful. The company uses AI constantly and values original expert judgment rather than AI output presented as one’s own. Employees start on a bench aligned with their expertise and may move across teams as they build a track record. The position offers output-based individual-contributor and management growth paths. It may be fully remote, though onsite work is preferred at the company’s Mission Bay waterfront office in San Francisco. In the US, the role is full-time W-2, with a guaranteed state-set base salary and OTE-based pay. The first four weeks are paid at full OTE during ramp-up, and full OTE is paid when quota cannot fairly apply. Outside the US, the role is contractor-based and paid on performance, with a monthly advance of 60% of OTE that is never clawed back; US benefits do not apply. The interview process includes an application, take-home project, technical interview, culture interview, offer, and background check after acceptance.
Requirements
- Professional experience working with medical images and experience handling thousands of studies.
- Qualifying professional background may include radiologist (including training outside the US, board-certified or equivalent), radiology resident or fellow, sonographer (ARDMS or equivalent), or 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 in the area of expertise.
- Ability to recognize image-quality problems such as artifacts, positioning issues, and parameter issues.
- Comfort using Python for interpretation work and running scripts.
- Daily use of popular AI agents, such as Claude Code, with judgment to distinguish technical correctness from clinical soundness.
Nice To Haves
- Imaging AI challenge experience, such as RSNA or MICCAI.
- Contributions to resources such as Radiopaedia.
- Familiarity with public imaging archives such as TCIA, MIMIC-CXR, UK Biobank, or OpenNeuro.
- ML engineering experience may be useful but is not required.
Responsibilities
- Determine the optimal answer for each medical-imaging case and build studies that test it.
- Assess whether scans were acquired properly and determine what the images show.
- Write questions frequently asked by radiologists or during medical-image interpretation.
- Choose the optimal answer format for the AI agent.
- Define what constitutes a passing result based on the objective.
- Review how AI agents handle cases, including missed findings, unsupported findings, and the soundness of their approach.
- Help determine what an AI agent needs to solve each objective.
Benefits
- Health, dental, and vision coverage fully covered for the employee and 50% covered for dependents (US only).
- 401(k) with no match (US only).
- 12 days of PTO to start, with more as the employee is promoted.
- Four-week ramp paid at full OTE.
- Full OTE pay when quota cannot fairly apply, such as during a new project, platform outage, or special project.
- Company outings most months and team off-sites.
- Catered lunch and dinner and unlimited snacks at the San Francisco office.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Entry Level
Education Level
No Education Listed
