ML Research Engineer / Scientist

Accountabilities

• Design, develop, train, and evaluate machine learning models capable of interpreting complete CT studies at the study level.
• Research foundation-model approaches for medical imaging, including 3D and volumetric learning at large scale.
• Develop and investigate vision-language models that connect medical images with the terminology and reporting patterns used by radiologists.
• Build models capable of identifying and prioritizing multiple urgent clinical findings simultaneously while maintaining safe and clinically appropriate operating points.
• Design and execute independent experiments, from hypothesis formation and architecture selection through training, evaluation, and analysis.
• Develop custom architectures, training pipelines, loss functions, and distributed training approaches using modern deep learning frameworks.
• Analyze model performance rigorously and establish reproducible evaluation methodologies suitable for clinically consequential AI systems.
• Work closely with fellowship-trained radiologists to understand clinical requirements, interpret results, and translate research findings into practical model improvements.
• Contribute to models and research that progress toward regulatory submissions, clinical deployment, and real-world patient use.
• Take ownership of research projects end to end and make informed decisions about which experiments and approaches are most likely to deliver meaningful improvements.
• Collaborate with ML and software engineering teams to move successful research from experimentation toward robust, deployable systems.

Requirements:

• Strong practical experience with modern machine learning and deep learning, particularly using PyTorch for custom architectures, training loops, and experimentation.
• Deep understanding of why machine learning architectures, objectives, optimization strategies, and training approaches work, rather than relying solely on existing implementations.
• Demonstrated ability to independently formulate hypotheses, design experiments, interpret results, and iterate toward better models.
• Strong understanding of rigorous experimentation, evaluation, reproducibility, and model validation.
• Experience working with large-scale datasets and distributed training environments is highly valuable.
• A strong interest in solving technically challenging problems where model performance and reliability have meaningful real-world consequences.
• Ability to work effectively with researchers, engineers, and clinical experts in a collaborative environment.
• Medical imaging, 3D computer vision, or volumetric-data experience is advantageous but not required.
• Experience with vision-language models or self-supervised learning is a plus.
• Familiarity with DICOM, CT imaging, radiology, or other medical-data formats and workflows is beneficial.
• A PhD, research publications, or a strong academic research background is a plus, but not a prerequisite.
• Prior medical-AI experience is not required; a willingness to learn clinical concepts directly from radiology experts is valued.
• Strong written and verbal communication skills and the ability to work independently in a fully remote environment.

Benefits:

• Fully remote position open to candidates worldwide.
• Location-flexible compensation with a cash-weighted base salary determined according to the local market in the country where you work.
• Specific compensation range for your location shared early in the hiring process.
• No equity included in international offers, with compensation structured transparently around local-market cash pay.
• Opportunity to work with a real-world CT dataset covering approximately 10 million patients, paired with radiology reports.
• Direct collaboration with fellowship-trained radiologists across areas including chest, body, MSK, neuro, and oncology.
• Opportunity to work on research that can progress from experimentation to FDA submissions, hospital deployments, and patient care within months.
• Exposure to large-scale foundation models, vision-language learning, distributed training, medical imaging, and clinically focused AI evaluation.
• High degree of ownership over research ideas, experiments, models, and technical direction.
• Opportunity to work alongside researchers and engineers with significant contributions to medical AI, open datasets, algorithms, and clinical AI systems.
• A small, research-oriented team where successful ideas can move quickly from research into production.