Director, AI Engineering - Acadia Pharmaceuticals Inc.

OpenTalent; Princeton (NJ); Hybrid; USD 180,000 - 260,000; Full time

Job summary

Acadia is committed to turning scientific promise into meaningful innovation that improves outcomes for underserved neurological and rare disease communities worldwide. The Director, AI Engineering leads the enterprise AI platform and agent ecosystem to enable scalable, secure, and impactful AI-powered solutions across the organization.

In partnership with IT, Cybersecurity, Enterprise Architecture, and business stakeholders, you will drive reusable templates, multi-agent orchestration, and

Responsibilities

• Architect and maintain the enterprise agent platform - the system through which agents are developed, published, versioned, and coordinated, using frameworks such as LangGraph, CrewAI, Anthropic Agents SDK, and OpenAI Agents SDK.
• Engineer multi-agent coordination patterns enabling agents with distinct roles, data access, and functional scope to collaborate reliably on complex, multi-step business workflows.
• Design and implement reusable agent templates, composable orchestration primitives, and agentic memory systems (short-term, long-term, episodic) that development teams and non-engineers can configure and deploy.
• Build human-in-the-loop (HITL) escalation workflows and clearly defined agent action boundaries to ensure appropriate human oversight at high-risk decision points.

Job description

Acadia is committed to turning scientific promise into meaningful innovation that makes the difference for underserved neurological and rare disease communities around the world. Our commercial portfolio includes the first and only FDA-approved treatments for Parkinson’s disease psychosis and Rett syndrome. We are developing the next wave of therapeutic advancements with a robust and diverse pipeline that includes mid- to late-stage programs in Alzheimer’s disease psychosis and Lewy body dementia psychosis, along with earlier-stage programs that address other underserved patient needs. At Acadia, we’re here to be their difference.

This hybrid position requires onsite attendance three days per week at one of our office locations in San Diego, CA, South San Francisco, CA, or Princeton, N J

Position Summary

The Director, AI Engineering is a strategic technology leader responsible for shaping and advancing the enterprise AI application and agent ecosystem. This role owns the AI platform and agent framework layer that enables scalable, secure, and impactful AI-powered solutions across the organization.

The successful candidate will drive the development of enterprise-grade capabilities including reusable agent templates, multi-agent orchestration, an internal agent publishing platform, self-service RAG solutions, tool and skill registries, prompt management, and evaluation and output validation frameworks. This leader will balance deep technical expertise with a passion for creating intuitive, low-code experiences that empower business users to independently build and deploy AI solutions.

Partnering closely with IT, Cybersecurity, Enterprise Architecture, and business stakeholders, the Director will ensure the AI platform remains secure, governed, scalable, and aligned with organizational objectives while delivering measurable business value and accelerating AI adoption across the enterprise.

Primary Responsibilities

• Architect and maintain the enterprise agent platform - the system through which agents are developed, published, versioned, and coordinated, using frameworks such as LangGraph, CrewAI, Anthropic Agents SDK, and OpenAI Agents SDK.
• Engineer multi-agent coordination patterns enabling agents with distinct roles, data access, and functional scope to collaborate reliably on complex, multi-step business workflows.
• Design and implement reusable agent templates, composable orchestration primitives, and agentic memory systems (short-term, long-term, episodic) that development teams and non-engineers can configure and deploy.
• Build human-in-the-loop (HITL) escalation workflows and clearly defined agent action boundaries to ensure appropriate human oversight at high-risk decision points.