New opportunity

Principal AI Scientist

Solventum · Bangalore, India

About this role

As a Principal AI Scientist at Solventum, you will lead high-impact technical innovation in healthcare information systems. This role involves driving technical breakthroughs, designing novel AI architectures, and integrating advanced AI into mission-critical healthcare products. Responsibilities include researching and prototyping autonomous, agent-based systems, advancing NLU for clinical text, and adapting LLMs for healthcare-specific regulatory and ethical requirements. You will serve as a technical authority for AI/ML strategy, ensuring models are explainable, robust, and compliant with HIPAA/HITRUST standards. You will also lead the full lifecycle of AI applications, from data exploration to production deployment, while mentoring the engineering team. Required Qualifications: PhD or Master’s in Computer Science, AI, Machine Learning, or a related field; 5+ years of experience in AI/ML with a portfolio of deployed models; proficiency in Python, R, and modern ML libraries; experience with GenAI, LLMs, transformer architectures, and AI frameworks (PyTorch, TensorFlow, Hugging Face, LangChain, RAG); hands-on experience with autonomous AI agents; and familiarity with AWS and MLOps. Preferred Skills: Experience with multi-modal AI research, knowledge graph integration, symbolic reasoning, and a proven track record of taking research from concept to production.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Agentic AIGenerative AINatural Language UnderstandingPythonRPyTorchTensorFlowHugging FaceLangChainRAGAWSMLOpsFHIRHL7Classical MLDeep LearningKnowledge Graphs

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Architecting Autonomous Agentic Systems
The role requires designing and prototyping agent-based systems for healthcare workflows. You must demonstrate how to build reasoning systems that automate complex tasks while maintaining strict human-in-the-loop safety standards and clinical accuracy.
Ensuring Regulatory Compliance and Model Explainability
As a technical authority, you must ensure all AI models meet HIPAA and HITRUST standards. You need to prioritize the development of explainable AI (XAI) frameworks that provide transparency for clinical decision-making and mitigate bias in sensitive healthcare data.
Bridging Research to Production
You are expected to lead the full lifecycle of AI applications. This requires balancing advanced research in LLMs and multi-modal AI with practical MLOps, ensuring that prototypes are scalable, secure, and maintainable in a production environment.
Clinical Data Integration and NLU
You must advance NLU capabilities for clinical text and integrate structured/unstructured data (FHIR, HL7). Success depends on your ability to create pipelines that process diverse healthcare data formats while maintaining high data integrity.

Likely challenges to think through

Balancing Innovation with Healthcare Rigor
The primary challenge is implementing cutting-edge GenAI and agentic frameworks within a highly regulated, risk-averse healthcare environment. Prepare by articulating how you manage the trade-off between model performance and the absolute necessity for clinical safety and regulatory adherence.
Managing Complex Data Heterogeneity
Healthcare data is notoriously fragmented across FHIR, HL7, and unstructured clinical notes. You may face difficulty in unifying these sources for model training; prepare by discussing your experience with data cleaning, normalization, and knowledge graph integration to improve model reasoning.
Technical Leadership in a Hybrid Environment
As a Principal Scientist, you must mentor teams and set strategy while remaining hands-on. You will likely face challenges in maintaining technical alignment across diverse engineering teams; prepare by highlighting your experience in creating architectural design documentation and conducting rigorous code reviews.

Practice your answers

Technical practice

Designing Autonomous Agentic Systems in Healthcare

What to coverExplain your approach to architecting agent-based systems that automate clinical workflows. Focus on how you maintain 'human-in-the-loop' safety standards while ensuring the agent's reasoning remains reliable and explainable.

RAG Implementation for Clinical Data

What to coverDescribe how you would build a Retrieval-Augmented Generation (RAG) pipeline to process unstructured clinical notes alongside structured FHIR/HL7 data. Discuss your strategy for ensuring data integrity and minimizing hallucinations in a high-stakes healthcare environment.

Balancing LLMs with Classical ML

What to coverThe role requires combining Classical ML (e.g., XGBoost) with Deep Learning and Generative AI. Provide an example of a scenario where a hybrid modeling approach is superior to using a single architecture, and how you manage the integration.

Ensuring Regulatory Compliance (HIPAA/HITRUST)

What to coverHow do you incorporate HIPAA and HITRUST requirements into the AI development lifecycle? Suggest how you design model evaluation frameworks to ensure compliance, data security, and privacy from the prototyping phase to production.

Advancing NLU for Clinical Text

What to coverDiscuss the challenges of applying NLU to clinical and administrative healthcare text. Explain how you would adapt transformer architectures to handle domain-specific terminology and improve accuracy in clinical decision-making tasks.

MLOps for Healthcare AI

What to coverDetail your experience in moving research prototypes into production on AWS. Focus on your strategy for monitoring model performance, detecting bias, and ensuring clinical accuracy over time in a production environment.

Knowledge Graph Integration

What to coverExplain the role of knowledge graphs and symbolic reasoning in enhancing the performance of LLMs. Suggest how you would integrate these with neural models to improve reasoning capabilities in complex healthcare workflows.

Multi-modal AI Research

What to coverGiven the preference for multi-modal experience, describe how you would design an architecture that processes text, images, and structured data simultaneously. Discuss the trade-offs in model complexity versus performance gains for clinical applications.

Behavioural practice

Bridging Research and Engineering

What to coverDescribe a situation where you had to transition a theoretical AI research project into a production-ready application. Detail the specific actions you took to align the research goals with engineering constraints and the final outcome.

Technical Leadership and Mentorship

What to coverProvide an example of how you have elevated an engineering team's capabilities through code reviews or architectural guidance. Focus on a specific instance where your mentorship improved the team's adherence to data science best practices.

Navigating Ethical Dilemmas in AI

What to coverShare an experience where you had to address an ethical concern or bias in an AI model. Explain the situation, the steps you took to mitigate the risk, and how you ensured the final model met safety and fairness standards.

Collaborating with Domain Experts

What to coverDescribe a project where you worked closely with non-technical domain experts (e.g., clinicians or healthcare administrators). How did you translate their requirements into technical specifications, and what was the impact on the final product?

Sharpen your knowledge

Topics to prepare

Healthcare Regulatory Compliance
Review HIPAA and HITRUST standards as they apply to AI model deployment. Exercise: Draft a checklist for auditing an LLM pipeline for PII leakage and bias, ensuring compliance with data privacy requirements in clinical settings.
Agentic AI Architectures
Understand the design of autonomous systems with human-in-the-loop safety protocols. Exercise: Design a conceptual architecture for an agent that automates a clinical workflow, focusing on decision-making logic and safety guardrails.
Hybrid AI Modeling
Study the integration of classical machine learning with deep learning and generative models. Exercise: Compare the use cases for XGBoost versus Transformer-based models in predicting patient outcomes, identifying when to use each or a hybrid approach.

Skills to revise

RAG and LangChain Implementation
Build a Retrieval-Augmented Generation pipeline using LangChain. Exercise: Create a prototype that queries a mock clinical dataset (FHIR/HL7 format) to answer specific medical questions, emphasizing retrieval accuracy and source citation.
MLOps for Healthcare
Practice the full lifecycle of model deployment on AWS. Exercise: Outline a CI/CD pipeline for a healthcare AI model that includes automated testing for model drift, fairness, and clinical accuracy before production release.
Knowledge Graph Integration
Explore how knowledge graphs enhance symbolic reasoning in AI. Exercise: Map a set of clinical entities and their relationships into a graph structure to support a reasoning task, demonstrating how this improves upon standard LLM outputs.

Your preparation checklist

Prepare a portfolio of production-grade AI deployments
Curate a summary of your past projects that highlights the full lifecycle from research to production. Focus on examples where you integrated LLMs or agentic systems into complex environments, specifically emphasizing how you handled data security, scalability, and MLOps practices.
Develop a framework for healthcare-specific AI compliance
Prepare to discuss your approach to balancing innovation with HIPAA/HITRUST requirements. Create a brief outline or mental model for how you ensure model explainability, bias detection, and clinical accuracy when deploying AI in mission-critical healthcare workflows.
Design a technical walkthrough for an agentic healthcare workflow
Draft a high-level architecture for an autonomous agent system that processes clinical data (e.g., FHIR/HL7) to assist in decision-making. Be ready to explain your choice of frameworks like LangChain or RAG and how you maintain 'human-in-the-loop' safety standards.
Formulate strategic questions regarding Solventum's AI roadmap
Prepare thoughtful questions about how the team balances symbolic reasoning with deep learning, the current challenges in scaling their multi-modal research, and how they prioritize technical debt versus new innovation in their existing healthcare information systems.

Availability has not yet been verified. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…