Data Scientist - Medical document analysis

Thank you for your interest in joining Solventum. Solventum is a new healthcare company with a long legacy of solving big challenges that improve lives and help healthcare professionals perform at their best. At Solventum, people are at the heart of every innovation we pursue. Guided by empathy, insight, and clinical intelligence, we collaborate with the best minds in healthcare to address our customers’ toughest challenges. While we continue updating the Solventum Careers Page and applicant materials, some documents may still reflect legacy branding. Please note that all listed roles are Solventum positions, and our Privacy Policy: https://www.solventum.com/en-us/home/legal/website-privacy-statement/applicant-privacy/ applies to any personal information you submit. As it was with 3M, at Solventum all qualified applicants will receive consideration for employment without regard to their race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or status as a protected veteran.

Job Description:

Data Scientist for Medical Document Analysis (Solventum)

 

3M Health Care is now Solventum

 

At Solventum, we enable better, smarter, safer healthcare to improve lives. As a new company with a long legacy of creating breakthrough solutions for our customers’ toughest challenges, we pioneer game-changing innovations at the intersection of health, material, and data science that change patients' lives for the better while enabling healthcare professionals to perform at their best. Because people, and their wellbeing, are at the heart of every scientific advancement we pursue.

 

We partner closely with the brightest minds in healthcare to ensure that every solution we create melds the latest technology with compassion and empathy. Because at Solventum, we never stop solving for you.

 

The Impact You’ll Make in this Role:

 

As a Data Scientist specializing in medical document analysis, you will work at the forefront of healthcare NLP. You will design, implement, and evaluate advanced natural language understanding (NLU) systems that interpret complex clinical text and support data-driven medical decision making. This includes developing models grounded in transformers, generative AI, and research-oriented neural architectures.

 

You will collaborate closely with clinical domain experts, fellow data scientists, ML engineers, and product teams to bring research prototypes into production systems. You will explore new modeling approaches, optimize training pipelines, and contribute to the long-term direction of Solventum’s deep learning portfolio.

 

You Will Make an Impact By:

• Scoping and designing NLP/NLU applications that support clinical understanding of medical documents and downstream healthcare workflows.

• Translating business and clinical requirements into mathematical and experimental criteria, ensuring models are measurable, reliable, and auditable.

• Processing, aligning, and enriching diverse structured and unstructured medical datasets for deep learning, LLMs, and agentic AI systems.

• Developing and evaluating deep learning models, including transformer-based architectures and neural architectures such as CNNs and attention mechanisms for capturing linguistic patterns in clinical text.

• Understanding and mitigating model weaknesses, including bias, drift, hallucination behavior, and robustness concerns.

• Designing experiments and error analyses that explain model behavior and enable clear communication with both technical and non-technical audiences.

• Deploying models into Solventum’s cloud environment, partnering with ML engineering to ensure reliability, observability, and compliance within regulated healthcare systems.

• Staying current with emerging research in NLP/LLMs, evaluating new methods and identifying opportunities to advance Solventum’s deep learning initiatives.

 

Your Skills and Expertise

To set you up for success in this role from day one, Solventum requires (at a minimum) the following qualifications:

• <