Director, Agent Intelligence - HYBRID ROLE

Vertex Inc.•Boston, MA

7d•Hybrid

About The Position

This is a Hybrid position requiring 3 days a week in our Boston office. We are seeking an experienced AI engineering leader to build and operate Vertex's Agent Intelligence layer: the memory, personas, skills, and flagship agent experiences that make every Vertex AI agent grounded, consistent, and effective. Vertex Pharmaceuticals is in a transformational period, and the AI team is at the center of our AI-first strategy, delivering solutions that empower executives, researchers, and business users to make faster, more confident decisions. As part of the Vertex Data team, you will lead the Agent Intelligence function within Knowledge & Grounding. You will own the four primitives that all Vertex agents are built from: agent memory (personal and organizational), personas as code, skills (atomic tools and higher-order compositions), and the flagship agents that put those primitives to work on Vertex's hardest use cases. As the Director of Agent Intelligence, you will build the reusable building blocks of agent intelligence at Vertex. Your work will multiply the velocity and quality of agent development across Vertex, ensuring every agent benefits from a common foundation of grounded, governed, and reusable intelligence.

Requirements

10+ years of experience in data engineering, AI/ML, or advanced analytics, with 3+ years specifically focused on NLP systems, AI agent development, or conversational AI.
Hands-on experience building production AI agents using frameworks like LangChain, LangGraph, or similar orchestration tools.
Strong experience with Snowflake (Cortex AI, Cortex Analyst, Cortex Search) and/or Databricks (Agent Bricks, AI Functions, Mosaic AI) in enterprise environments.
Ability to work across the full AI stack, including data preparation, prompt engineering, agent orchestration, evaluation/testing, deployment, and monitoring.
Direct experience designing how agents remember, including session state, long-horizon memory, retrieval over conversational history, and the data model behind it.
Track record of operating AI systems in production with high availability, monitoring, and continuous improvement.
Proven ability to lead technical teams, communicate with executive stakeholders, and translate business needs into AI solutions.

Nice To Haves

Experience in pharmaceutical, biotech, or life sciences data environments. Familiarity with clinical, regulatory, and commercial data domains is a plus.
Experience designing reusable agent abstractions such as persona frameworks, skill registries, or tool catalogs.
Familiarity with evaluation tooling such as Braintrust, LangSmith, Mosaic AI Agent Evaluation, Arize, or Ragas.
Experience with model fine-tuning, distillation, or post-training for domain adaptation.

Responsibilities

Agent Memory. Design and operate Vertex's agent memory layer, spanning personal memory (per-user, per-session continuity) and organizational memory (shared institutional knowledge agents accumulate and reuse), including retention, retrieval, conflict resolution, and classification policies.
Personas as Code. Define and steward the persona system that gives Vertex agents their voice, remit, and boundaries, including versioning, review, composition, and deployment workflows.
Skills Catalog & Compositions. Curate the catalog of atomic tools agents can call (retrieval, analytics, action) and the higher-order skill compositions that combine them into persona-specific workflows, with clear contracts, registry, and quality standards.
Flagship Agent Development. Design, build, and deploy custom AI agents for Vertex's highest-value enterprise use cases using Snowflake Cortex AI, Databricks AI, and agent orchestration frameworks, using these builds to harden the underlying primitives.
Enterprise Agent Standards. Define and maintain the enterprise standards for agent orchestration, prompt libraries, persona definitions, tool-use patterns, and response quality that all teams must follow, regardless of builder.
Experience Consistency & UX Standards. Define the quality bar for conversational UX across the enterprise (tone/persona, citation patterns, error handling, escalation paths, transparency) and drive consistency across agents.
Prompt & Skill Library Stewardship. Curate and govern reusable prompt and skill assets (system prompts, persona templates, tool instructions, conversation patterns) with documentation and contribution workflows.