
Associate Director, AI Foundations Engineering (3 Openings)
Required skills
- agentic AI platform
- large language models
- agentic system
- AWS
- Amazon Bedrock
- Bedrock AgentCore
- Strands Agents
- LangGraph
- LlamaIndex
- CrewAI
Role summary
The Associate Director, AI Foundations Engineering will lead the engineering strategy, design, development, deployment, and optimization of an agentic AI platform. This role focuses on building scalable, secure, and reliable AI systems, partnering with cross-functional teams to define roadmaps and establish best practices for AI engineering. The ideal candidate has deep expertise in AI engineering, platform architecture, and production operations, with a proven ability to translate AI capabilities into business solutions. Responsibilities include overseeing AI engineering operations, guiding model integration, and leading high-performing engineering teams.
Job Description Summary
#LI-RemoteThe Associate Director, AI Foundations Engineering will lead the engineering strategy, design, development, deployment, and ongoing optimization of our agentic AI platform and solutions. This role is responsible for building scalable, secure, and reliable AI-driven systems that transform existing products, processes, and workflows into intelligent, agentic AI experiences. As a key technical leader, the Associate Director will partner closely with product, data, architecture, security, and business teams to define platform capabilities, drive implementation excellence, and operationalize continuous improvement across the AI lifecycle. The ideal candidate brings deep expertise in AI engineering, platform architecture, and production operations, along with a strong track record of translating emerging AI capabilities into practical, high-impact business solutions.
This position can be based remotely anywhere in the U.S. (there may be some
restrictions based on legal entity). Please note that this role would not provide
relocation as a result. The expectation of working hours and travel (domestic and/or international) will be defined by the Hiring Manager.
There are 3 positions available.
Job Description
Key Responsibilities:
- Lead the architecture, design, and engineering delivery of the organization's agentic AI platform and related solutions.
- Drive the transformation of existing products, processes, and workflows into scalable agentic AI-enabled capabilities.
- Oversee end-to-end AI engineering operations, including development, deployment, monitoring, reliability, and continuous improvement.
- Partner with cross-functional stakeholders to define technical roadmaps, platform standards, and solution priorities aligned to business goals.
- Establish engineering best practices for AI system performance, scalability, security, governance, and maintainability.
- Guide the evaluation, integration, and optimization of AI models, orchestration frameworks, and supporting platform components.
- Build, mentor, and lead high-performing engineering teams while fostering innovation, accountability, and technical excellence.
- Identify opportunities to accelerate value delivery through reusable AI services, automation, and operational efficiencies.
Essential Requirements:
- Bachelor's degree in computer science, engineering, or a related field, or equivalent practical experience. An advanced degree is a plus but not required.
- 5+ years of professional software engineering experience, including at least 1–2 years building applications on large language models, with at least one agentic system deployed beyond a demo.
- Hands-on experience building AI agents on AWS. This includes Amazon Bedrock (models, Agents, Knowledge Bases, Guardrails) and Bedrock AgentCore or comparable approaches for agent runtime, memory, tool access, and identity.
- Proficiency with at least one agent framework, such as Strands Agents, LangGraph, LlamaIndex, CrewAI, or a custom orchestration layer. You should be able to explain when a simple workflow beats a multi-agent design.
- Strong Python skills plus experience with AWS services commonly used in agent architectures: Lambda, Step Functions, ECS/EKS, API Gateway, DynamoDB, S3, and IAM.
- Experience with retrieval-augmented generation (RAG), including chunking strategies, embeddings, vector stores (OpenSearch, pgvector, or similar), and improving retrieval quality.
- Experience designing tool use and function calling so agents can connect to APIs, databases, and enterprise systems. Familiarity with Model Context Protocol (MCP) is preferred.
- A practical approach to evaluation and observability: test sets for agent behavior, LLM-as-judge or human review loops, tracing (CloudWatch, OpenTelemetry, LangSmith, or similar), and monitoring cost and latency.
- Ability to take an ambiguous business problem to a working prototype quickly, then harden it for real users with security, guardrails, error handling, and CI/CD using CDK or Terraform.
- Clear communication of technical tradeoffs to product and business partners, and comfort working in a small, fast-moving team.
Novartis Compensation Summary:
The salary for this position is expected to range between $145,600 and $270,400 per year.
The final salary offered is determined based on factors like, but not limited to, relevant skills and experience, and upon joining Novartis will be reviewed periodically. Novartis may change the published salary range based on company and market factors.
Your compensation will include a performance-based cash incentive and, depending on the level of the role, eligibility to be considered for annual equity awards.
US-based eligible employees will receive a comprehensive benefits package that includes health, life and disability benefits, a 401(k) with company contribution and match, and a variety of other benefits. In addition, employees are eligible for a generous time off package including vacation, personal days, holidays and other leaves.
EEO Statement:
The Novartis Group of Companies are Equal Opportunity Employers. We do not discriminate in recruitment, hiring, training, promotion or other employment practices for reasons of race, color, religion, sex, national origin, age, sexual orientation, gender identity or expression, marital or veteran status, disability, or any other legally protected status.
Accessibility and reasonable accommodations
The Novartis Group of Companies are committed to working with and providing reasonable accommodation to individuals with disabilities. If, because of a medical condition or disability, you need a reasonable accommodation for any part of the application process, or to perform the essential functions of a position, please send an e-mail to us.reasonableaccommodations@novartis.com or call +1(877)395-2339 and let us know the nature of your request and your contact information. Please include the job requisition number in your message.
Salary Range
$145,600.00 - $270,400.00Skills Desired
Cross-Functional Collaboration, Customer Engagement, Customer Insights, Data Analytics, Digital Marketing, Market Research, Media Campaigns, Product Marketing, Stakeholder Engagement, Stakeholder Management, Team Leadership, Waterfall ModelSample Novartis interview questions
- 1
Merge a New Interval Merge a new interval into a list of non-overlapping intervals. Input: intervals = [[1,2],[3,5],[6,7],[8,10],[12,16]], newInterval = [4,8] Output: [[1,2],[3,10],[12,16]] Explanation: The new interval overlaps with [3,5], [6,7], and [8,10], merging them all into the unified block [3,10].
codingmedium - 2
Sort a String by Frequency Sort a string by the frequency of its characters. Input: s = "cccaaa" Output: "cccaaa" or "aaaccc" Explanation: Both 'c' and 'a' appear three times, so sorting them by descending frequency keeps them grouped equally.
codingmedium - 3
Diameter of a Binary Tree Find the diameter of a binary tree. Input: root = [1,2] Output: 1 Explanation: The longest path is simply the single edge connecting the root node to its only child.
codingmedium - 4
Iterator over a Binary Search Tree Implement an iterator over a binary search tree. Input: root = [7,3,15,null,null,9,20], calls: next(), hasNext(), next() Output: 3, TRUE, 7 Explanation: The iterator yields the smallest value (3), confirms more nodes exist, then properly yields the next in-order value (7).
codingmedium - 5
Minimum Path Sum Find the minimum path sum from top-left to bottom-right in a grid. Input: grid = [[1,2,3],[4,5,6]] Output: 12 Explanation: The optimal path moves along 1 -> 2 -> 3 -> 6, resulting in a minimum accumulated sum of 12.
codingmedium
Get started to unlock a personalized interview prep pack for this role.
