Johnson & Johnson is seeking an AI/ML Intern, Computer Vision to contribute to foundational research and development for large-scale, multi-modal visual models applied to medical and clinical imaging. You'll work on modular architectures, predictive and alignment-based methods for contextual understanding, and improvements to representation-learning pipelines.
What You'll Do
- Design, implement, and evaluate scalable modular model architectures that allow specialization and efficient use of computation.
- Develop and test methods that learn richer contextual and temporal representations by predicting or aligning different views, frames, or modalities.
- Improve representation-learning pipelines by experimenting with data preparation strategies, augmentation approaches, training schedules, and hyperparameter settings to increase robustness across modalities.
- Build reproducible training and evaluation workflows and run experiments at scale; maintain clear experiment logs and analyses.
- Measure model effectiveness on clinically relevant downstream tasks (e.g., classification, detection, segmentation, retrieval, temporal reasoning) and produce comparison reports and ablation studies.
- Collaborate with data engineers, clinicians, and researchers to curate and prepare datasets while following privacy and governance requirements.
- Produce well-documented code, experiment artifacts, internal reports, and, where appropriate, contribute to technical write-ups or presentations.
What We're Looking For
- Currently pursuing or recently completed a Bachelor’s, Master’s, or PhD in Computer Science, Engineering, Applied Mathematics, or a related field.
- Strong programming ability (Python) and experience with common machine learning libraries.
- Solid understanding of machine learning and computer vision fundamentals and of how to train and evaluate models.
- Experience running experiments, tracking results, and performing basic troubleshooting and analysis.
- Strong written and verbal communication skills.
- Permanently authorized to work in the U.S., must not require sponsorship of an employment visa (e.g., H-1B or green card) at the time of application or in the future.
Nice to Have
- Prior research or project experience related to modular model design, predictive/alignment methods for representation learning, or representation learning using unlabeled data.
- Experience working with medical imaging or multi-modal visual data (including video) and familiarity with common preprocessing challenges.
- Experience with training models at scale and with experiment management practices.
- Understanding of clinical evaluation metrics and concerns around generalization and robustness in medical imaging.
- Publications, open-source contributions, or a portfolio demonstrating relevant work.
Technical Stack
- Python, common machine learning libraries
Benefits & Compensation
- Compensation: $23.00 per hour to $51.50 per hour (expected range)
- Eligible to participate in Company sponsored employee medical benefits.
- Eligible for sick time benefits: up to 40 hours per calendar year; for employees who reside in the State of Washington, up to 56 hours per calendar year.
- Eligible to participate in the Company’s consolidated retirement plan (pension).
Work Mode
This is a local-city role based in one of the following locations: New Brunswick, New Jersey, United States of America; Raritan, New Jersey, United States of America; San Diego, California, United States of America; Spring House, Pennsylvania, United States of America; Washington, District of Columbia, United States of America.
Johnson & Johnson is an Equal Opportunity Employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, age, national origin, disability, protected veteran status or other characteristics protected by federal, state or local law.
