Junior Machine Learning Engineer - Medical Imaging

Junior Machine Learning Engineer – Medical Imaging

Holobeam Technologies | Bethpage, NY

Holobeam Technologies is developing advanced medical imaging technologies designed to improve how medical images are processed, analyzed, and used in clinical care.

We are looking for a Junior Machine Learning Engineer – Medical Imaging who wants to do more than maintain existing models. This is a hands-on research and development role for someone interested in building and testing new approaches in medical imaging, computer vision, and machine learning.

You will work with real medical imaging datasets, develop and evaluate ML models, investigate new research ideas, and help turn promising concepts into reproducible imaging pipelines.

As you grow in the role, we want you to be able to take ownership of a research direction understand the relevant literature, reproduce or modify published approaches, design experiments, train models, investigate failures, and determine what actually works.

What You’ll Do

• Review and validate medical imaging datasets, metadata, annotations, and model outputs.
• Develop, train, benchmark, and evaluate computer vision and medical-image models using Python and PyTorch.
• Work with X-ray, CT and other medical imaging data, including DICOM studies, annotations, masks, and volumetric data.
• Build reproducible preprocessing, training, inference, and evaluation pipelines.
• Implement and evaluate methods from research papers and investigate ways to improve them.
• Design controlled experiments to compare architectures, preprocessing methods, loss functions, training strategies, and datasets.
• Perform detailed error and failure analysis rather than relying only on aggregate model metrics.
• Identify problems involving data quality, preprocessing, labeling, distribution shift, and model generalization.
• Develop automated quality-control and data-validation tools.
• Document experiments, methods, decisions, failures, and results so work can be reproduced.
• Work directly with technical and clinical team members to determine whether experimental results are meaningful.
• Take increasing ownership of an individual research and model-training direction as your experience grows.

What We’re Looking For

• Strong Python programming skills.
• Hands-on experience with PyTorch.
• Experience training and evaluating computer vision or deep-learning models.
• Understanding of common architectures used for tasks such as classification, detection, or image analysis.
• Experience working with image data and image-processing pipelines.
• Ability to independently debug problems involving data, code, training, and model outputs.
• Understanding of proper experimental design, including train/validation/test separation, evaluation metrics, and error analysis.
• Ability to read technical papers and translate an algorithm or research approach into working code.
• Strong curiosity and willingness to investigate why an approach works or fails, rather than simply running an existing pipeline.
• Ability to communicate experimental results clearly and document work reproducibly.

Particularly Valuable Experience

You do not need to have everything below. Experience in one or more of these areas will make you especially interesting to us:

• Medical imaging: X-ray, CT, MRI, PET, ultrasound, OCT, or CBCT
• DICOM and medical-image preprocessing
• Image reconstruction or inverse problems
• Image registration
• Diffusion models or other generative models
• 3D computer vision or volumetric imaging
• Scientific image processing
• GPU training and CUDA
• Reproducing or extending published ML research
• Research resulting in a thesis, publication, conference paper, or substantial independent project

Who Might Be a Great Fit

You might be finishing a bachelor’s or master’s degree, working toward or recently completing a PhD, or already have early-career industry/research experience.

We care more about what you have actually built and investigated than the exact title of your degree.

For example, we’d be interested in someone who can say:

“I implemented a segmentation paper and discovered why it didn’t generalize to my dataset.”

“I built a preprocessing and training pipeline for CT or MRI data.”

“I trained a model, analyzed its failure cases, changed the approach, and demonstrated why the new version was better.”

“I found a problem with the dataset or its preprocessing that was causing the model to fail and fixed it.”

“I took a research idea from literature review through experimentation and evaluation.”

That kind of technical ownership and curiosity is important to us.

What Makes This Role Different

This is not primarily a data-labeling, data-cleaning, or model-monitoring position.

Those activities are part of developing reliable medical-imaging AI, but our goal is to develop someone who can become an independent ML researcher/engineer capable of owning an experimental direction.

You will have the opportunity to work on difficult imaging problems where the answer may not already be known. You’ll be expected to experiment, question assumptions, learn new methods and help determine which approaches deserve further development.

About Holobeam

Holobeam Technologies is a medical technology company developing advanced imaging technologies with the goal of improving clinical care and helping save lives.

Our team works at the intersection of medical imaging, computer vision, machine learning, image processing and emerging imaging technologies.

We offer competitive salaries commensurate with experience, plus extremely generous health care plan, as well as 401K and stock options.

Location: Bethpage, Long Island, New York

Apply on LinkedIn

View email address on click.appcast.io for specific questions.