Associate Director, AI​/ML Engineering

Please note that this position is based in San Diego, CA, South San Francisco, CA, or Princeton, NJ. Acadia's hybrid model requires this role to work in our office three days per week on average. Position Summary

The Associate Director, AI/ML Engineering serves as a hands‑on technical leader driving the design, architecture, and delivery of Generative AI and agentic AI solutions across the enterprise. This role builds scalable multi‑agent systems, connects AI solutions to enterprise data and tools, and ensures safe, reliable deployment through robust evaluation and guardrail frameworks. The position also applies strong machine learning and foundation model expertise to deliver high‑impact use cases within a regulated biopharmaceutical environment.

• Design, build, and deploy agentic AI workflows that automate and transform complex business processes, leveraging multi‑agent orchestration frameworks (e.g., Lang Graph, Auto Gen, CrewAI, or equivalent).

• Architect and implement MCP servers to expose enterprise tools, APIs, and data sources as standardized capabilities consumable by AI agents.

• Connect multi‑agent systems to enterprise databases, internal APIs, and MCP servers to enable grounded, context‑aware, and action‑oriented AI solutions.

• Partner cross‑functionally with internal teams to define data contracts, lineage standards, and quality thresholds required for AI/ML use cases.

• Design and implement agentic memory systems (short‑term, long‑term, episodic) and planning/reasoning loops to support reliable autonomous task execution.

• Evaluate agentic system performance across accuracy, reliability, latency, cost, and safety dimensions using structured benchmarks and red‑team methods.

• Build and maintain guardrail frameworks (input/output filtering, content moderation, policy enforcement, hallucination detection) to ensure the safety, compliance, and trustworthiness of GenAI and agentic solutions.

• Develop retrieval‑augmented generation (RAG) pipelines, including chunking strategies, embedding models, vector store selection, and retrieval optimization for enterprise knowledge bases.

• Apply prompt engineering, few‑shot learning, and fine‑tuning techniques to adapt foundation models for domain‑specific pharma use cases.

• Design, develop, validate, and deploy traditional machine learning models (classification, regression, clustering, time‑series, survival analysis) to address structured business problems.

• Build and maintain end‑to‑end ML pipelines adhering to LLM Ops / ML Ops standards including model registry, evaluation benchmarks, prompt/version control, observability, and rollback procedures.

• Experience in working with real‑world data (RWD), claims data, EHR data, Clinical Study data, translational and biological data and the corresponding databases is a plus.

• Other responsibilities as assigned.

• Master’s or PhD in Machine Learning, Computer Science, Data Science, Information Systems, or a related quantitative discipline
• Minimum of 7 years of experience in AI/ML engineering, including at least 3 years of hands‑on experience with Generative AI and agentic AI systems
• Expertise in multi‑agent frameworks such as Lang Graph, Auto Gen, CrewAI, Semantic Kernel, or similar technologies
• Experience building MCP servers and integrating AI systems with enterprise data sources, APIs, and tools
• Strong experience in RAG pipeline development, embedding models, and vector database technologies
• Proficiency in Python and machine learning frameworks such as PyTorch, Tensor Flow, scikit‑learn, and Hugging Face
• Experience implementing ML Ops or LLM Ops practices, including model lifecycle management, evaluation, and deployment
• Ability to travel domestically and internationally as required

This role involves regular standing, walking, sitting, and the use of hands for handling or operating equipment. The employee may also need to reach, climb, balance, stoop, kneel, crouch, and maintain visual, verbal, and auditory communication in a standard office environment and while working independently from remote locations. The employee must occasionally lift and/or move up…