Optum is a global organization that delivers care, aided by technology to help millions of people live healthier lives. The work you do with our team will directly improve health outcomes by connecting people with the care, pharmacy benefits, data and resources they need to feel their best. Here, you will find a culture guided by inclusion, talented peers, comprehensive advantages and career development opportunities. Come make an impact on the communities we serve as you help us advance health optimization on a global scale. Join us to start Caring. Connecting. Growing together.

We're looking for a hands-on AI/ML Applied Scientist to drive experimentation, rapid prototyping, and solution innovation across AI, Generative AI, and Agentic AI technologies. You will work closely with business stakeholders, product teams, and engineering organizations to translate complex business challenges into practical AI solutions. The role focuses on proof-of-concept development, technology exploration, GenAI experimentation, and enabling the successful transition of validated solutions into production environments.

Unlike traditional data science roles focused solely on experimentation, this role is accountable for developing production-ready AI solution blueprints and ensuring successful engineering adoption. Success will be measured not only by innovation outcomes but also by the percentage of AI solutions that successfully progress from proof-of-concept to production deployment and deliver measurable business value. - Primary Responsibilities :

- Translate business problems into AI/ML, Generative AI, and Agentic AI solution approaches

- Conduct hands-on experimentation using machine learning, Generative AI, Agentic AI, and emerging AI technologies

- Design, build, and validate proof-of-concepts (POCs)



and prototypes to assess technical feasibility, business value, scalability, and operational readiness

- Develop production-oriented POCs that establish implementation patterns, reusable assets, architecture guidance, deployment approaches, and operational considerations required for enterprise adoption

- Create reusable prompts, workflows, evaluation frameworks, reference architectures, solution accelerators, and implementation assets for broader organizational adoption

- Drive successful transition of validated POCs into production by partnering closely with engineering teams to ensure solutions are scalable, maintainable, secure, and aligned with enterprise architecture standards

- Develop implementation-ready artifacts including reusable code components, prompt libraries, workflow templates, deployment recommendations, evaluation methodologies, and technical documentation to accelerate engineering adoption

- Own the technical readiness of AI solutions by proactively identifying scalability constraints, operational dependencies, implementation risks, and mitigation strategies during experimentation

- Structured evaluation and benchmarking

- Iterative model refinement

- Experiment tracking and documentation

- Performance and cost optimization

- Document learnings, experimentation results, architectural recommendations, and reusable solution assets

- Develop Generative AI solutions leveraging:

- Retrieval-Augmented Generation (RAG) architectures





- Prompt engineering and optimization techniques

- Vector databases and semantic retrieval frameworks

- AI evaluation and guardrails

- Build and evaluate Agentic AI workflows, including:

- Tool integration and orchestration

- Multi-step reasoning and planning

- Multi-agent collaboration patterns

- Autonomous and semi-autonomous workflows

- Evaluate emerging AI frameworks, platforms, and technology stacks to identify opportunities for innovation, standardization, and enterprise adoption

- Support development and adoption of AI accelerators, reusable frameworks, and best practices across teams

- Optimize early-stage solution cost efficiency through:

- Token usage awareness and optimization

- Prompt tuning and response management

- Model selection based on use-case requirements and cost-performance targets

- Cost-performance tradeoff analysis

- Collaborate with business, product, architecture, and engineering teams to clarify requirements and align solutions with measurable business outcomes

- Communicate experimentation results, trade-offs, recommendations, implementation considerations, and business impact to technical and non-technical stakeholders

- Accelerate organizational AI adoption by reducing the cycle time from experimentation to production deployment through repeatable patterns and reusable assets

- Measure success through:

- Quality and business impact of AI/ML, GenAI, and Agentic AI POCs

- Production readiness of delivered solutions

- Percentage of POCs successfully adopted and deployed into production

- Adoption of reusable accelerators, prompts, workflows, and reference architectures

- Reduction in experimentation-to-production cycle time

- Delivery of me

📌 Senior Ai/ml Engineer - Llm, Rag, Agentic Ai (Bengaluru)
🏢 UnitedHealth Group
📍 Bengaluru