AI Engineer

Note: APA operates on a 37.5-hour work week with one-hour lunch breaks each day. APA sets salaries based on several considerations, including years of relevant experience, level of education, and previous staff and/or governance experience at APA.

The AI Engineer role designs, builds, and operates agentic AI and retrieval-augmented generation (RAG) services that automate multi-step business workflows safely and reliably. The position partners with engineering and business stakeholders to deliver deployable, observable AI systems with measurable quality, strong guardrails, and ongoing iteration as tools and best practices evolve.

• Bachelor's degree in Computer Science, Engineering, Information Systems, or a related field; equivalent practical experience considered.
• 5–10 years of software engineering, including 1 to 2 years of building production grade Generative AI solutions (agentic applications and/or RAG).
• Hands‑on experience taking both agentic applications (multi‑step, tool‑calling) and RAG systems (retrieval → generation) to production, with deep expertise in at least one and solid working knowledge of the other.
• Built and shipped production services and APIs on AWS, with attention to reliability, latency, and cost.
• Comfortable with full‑stack delivery using Angular and Node/Type Script to expose AI capabilities to users.
• Familiar with CI/CD and standard operational practices (automated builds/releases, monitoring, alerting).

• Advanced Python for backend/AI service development, plus Angular/Node (Type Script) for full‑stack delivery.
• Practical experience with at least one agent framework (e.g., Lang Chain, Lang Graph, CrewAI) and safe tool calling design (clear tool boundaries, permissions, validation).
• Solid understanding of RAG: chunking, embeddings, retrieval, and basic reranking, with at least one vector store (e.g., Open Search, Pinecone, pgvector).
• Ability to evaluate and improve AI systems using practical quality signals, agent task success and to catch regressions.
• Working knowledge of AWS deployment (serverless and/or containers) with logging, monitoring, and alerting.
• Familiarity with responsible/secure AI basics: least privilege, data access controls, and guardrails.
• Git‑based source control and CI/CD, with reliable release and rollback practices.
• Clear communication and a collaborative, delivery‑oriented mindset.

• Design and build agentic AI applications that carry out multi‑step tasks, with explicit tool boundaries, permissions, and safety controls for predictable behavior.
• Build and maintain RAG solutions end‑to‑end (ingestion, embeddings, retrieval, generation), tuned for relevance, latency, and reliability.
• Define practical quality metrics for agents (task success, tool‑call correctness) and retrieval (relevance, coverage) and iterate on the results.
• Integrate tools and data sources via MCP or similar standard interfaces for secure, reusable access.
• Deliver Angular/Node interfaces and APIs that make agentic and RAG capabilities usable by business teams.
• Deploy and operate AI services on AWS (e.g., Lambda, ECS/EKS) with sound logging, monitoring, and error handling (retries, fallbacks, guardrails).
• Enable workflow automation and integration by identifying high‑value repetitive processes and delivering AI‑driven automation layers that integrate into enterprise systems and platforms.
• Create and maintain technical documentation for architectures, interfaces, evaluation approaches, and operational runbooks to support maintainability and auditability.
• Collaborate cross‑functionally with product/business stakeholders to translate workflows into well‑scoped deliverables and to validate outcomes against measurable success…