Associate Director, AI Engineering

Deliver robust, tested and supportable systems from the team's working prototypes to R&D scientists. The Associate Director is the team's senior full-stack engineer and architect. The role owns the engineering standards, development and deployment patterns, and production readiness for every application the team ships, and decides how software built with agentic tools by scientists and colleagues becomes durable, validation-ready systems. Its decisions set how quickly prototypes become products, how reliable those products are, and whether they are ready for validation. The role partners with a peer Associate Director, AI Platform & Enablement, who turns proven team patterns into shared platform capabilities over time. This is a hands-on individual-contributor role: the Associate Director designs, builds, debugs and reviews production software, working in agentic development tools (e.g., Claude Code, Codex) throughout the day, alongside validated-system work that follows its own procedures.

Responsibilities:

- Own engineering standards and production readiness for every application the team ships. Select and maintain the team's standard languages, frameworks and supported versions based on security, maintainability, enterprise compatibility and what AI-assisted development builds most reliably. Make architecture, release and incident-response decisions without escalation, within agreed governance, security and quality requirements.

- Take prototypes built by scientists and colleagues with agentic tools to production grade: refactor, strengthen and productize them with sound architecture, tests, observability, security, documentation and operability. Assess prototypes against agreed engineering standards and work with stakeholders to recommend and implement retain, rebuild or retire decisions.

- Own application-to-platform integration: the production concerns that recur across applications, such as authentication and single sign-on, secrets, networking, deployment and monitoring. Solve them for the team, build them once into reusable patterns, and work with the AI Platform & Enablement Associate Director and enterprise platform teams to move proven patterns into the platform.

- Establish and maintain the team's standard development and deployment patterns: CI/CD pipelines, infrastructure as code, environments, starter templates, monitoring and operational runbooks for the team's applications and agents.

- Own code-review standards and the merge gate for the team's repositories, reviewing and approving changes from scientists, colleagues, contractors and embedded delivery engineers, none of whom report to the role. Run design and architecture reviews before applications enter shared environments, and set the review standard for AI-generated code.

- Design and build full-stack, cloud-native systems: APIs, orchestration and workflow services, data and vector stores, identity and access, and scientist-facing front ends, favoring managed services and simple, maintainable designs.

- Partner with Quality and computer system validation colleagues to define and implement risk-based engineering and validation practices for LLM and agent-based applications, and build the evidence case for them. Practices may include requirements traceability, model and configuration change control, regression evaluations, controlled releases and auditable records. Quality retains decision authority.

- Work directly with R&D scientists and the team's scientists to define software requirements, review their code and build engineering capability through pairing. Lead the productization of the team's flagship applications, directing embedded delivery or professional-services engineers and coaching colleagues on engineering practice.

- Gain acceptance from platform teams and Cyber for the team's deployment and access patterns. Implement application security practice in line with Cyber standards, and ensure the team's applications and delivery practice comply with the company's quality management system, security guidelines and IT policies.

- Own the production-support approach for the team's applications, including triage, escalation, incident follow-up and engineering fixes, with support responsibilities agreed across the team and platform partners.

- Bring production-readiness and risk assessments to senior leadership, with recommendations on build versus buy and on which prototypes to harden or retire.

Qualifications:

• Bachelor's degree in a physical, life or quantitative science or in engineering, or equivalent experience; advanced degree preferred.
• 10+ years building and operating production software, with a substantial share as a hands-on architect or principal-level engineer who thinks in systems.

• Experience building or operating software in regulated, validated environments (e.g., GxP, computer system validation, IQ/OQ/PQ, requirements traceability) required.

• Demonstrated, sustained daily use of agentic development tools (e.g., Claude Code, Codex) to design, build and ship software, with full ownership of generated code, evidenced by delivered software, professional or personal.

• Experience owning CI/CD, infrastructure as code and release engineering with a globally distributed team: pipelines, environments, deployment and rollback, and observability.

• Track record of taking inherited or prototype code to production grade: refactoring, test coverage, security hardening, performance and operability.

• Experience solving the production concerns that sit between applications and platforms (e.g., authentication and SSO, secrets, networking, access control) and turning them into reusable patterns.

• Fluency in at least one production language and the judgment to direct agents across a polyglot stack (the team's is Python, TypeScript and PostgreSQL).

• Deep cloud-native experience, AWS preferred (equivalent Azure or GCP depth considered): containers and serverless, workflow orchestration, IAM and least-privilege design, managed data stores.

• Hands-on LLM application patterns (RAG, vector search, multi-stage pipelines) preferred.

• Experience working directly with research scientists preferred.