Principal AI Engineer - Context - Agents and Context
About this role
To ensure AI agents can reliably reason over enterprise information, Elastic needs someone to own the production improvement loop for the Context Engine. The Context Engine team builds the knowledge layer that allows AI agents to interact with enterprise data stored in Elasticsearch, processing information from diverse sources into a structured AI Index.
As a Principal AI Engineer based in Ireland, this hire will lead the end-to-end improvement of how agents, automations, and skills behave in production. Responsibilities include designing evaluation frameworks, establishing robust telemetry, and ensuring safe changes can be shipped without breaking customer integrations. The ideal candidate brings extensive software engineering experience with AI-driven products, a pragmatic approach to eval-driven development, and strong backend capabilities in TypeScript or Python.
What you will do
- Manage the production improvement loop for the Context Engine by analyzing extraction automations, retrieval tools, and memory behavior through telemetry and offline evaluations.
- Establish safety protocols for iterating on agents and skills, including prompt versioning, regression coverage, staged evaluations, and release guardrails.
- Design comprehensive telemetry systems to capture agent traces, tool calls, and retrieval data that feed directly into Elasticsearch and evaluation dashboards.
- Collaborate with the data science team to define evaluation strategies, including golden datasets and quality, latency, and cost gates.
- Drive engineering standards across the team through code and design reviews, technical proposals, and mentorship in eval-driven development.
- Build and maintain public APIs and evolving data models while ensuring seamless backward compatibility for external consumers.
What you need
- Possess 10 or more years of software engineering experience, including recent work shipping and operating AI-driven products on active production traffic.
- Demonstrate a proven track record of eval-driven product improvement by diagnosing LLM behavior, designing evaluations, and measuring fixes.
- Bring direct experience building stateful agents with memory and safely iterating on prompts, skills, and tool behavior in production environments.
- Show familiarity with MCP, encompassing the exposure of public MCP servers and tools.
- Have hands-on experience designing telemetry for AI systems to inform engineering and product decisions.
- Exhibit strong backend engineering proficiency in either Python or TypeScript for building APIs, stateful workflows, and data pipelines.
- Demonstrate experience running end-to-end product experiments, from instrumentation to interpreting unattended execution results.
Nice to have
- Experience with agent frameworks such as LangGraph, CrewAI, Claude Agent SDK, or similar tools
- Agentic retrieval experience, including knowledge representation
- Practical Elasticsearch experience
Benefits
- Competitive pay based on work done
- Health coverage for you and your family in many locations
- Flexible locations and schedules for many roles
- Generous number of vacation days each year
- Charitable donation matching up to $2000 or local currency equivalent
- Up to 40 hours each year for volunteer projects
- Minimum of 16 weeks of parental leave
Skills and tech stack
Questions candidates ask
What is the remote and office policy for this role?
This is a hybrid role that offers the ability to craft your calendar with flexible locations and schedules for many roles at Elastic.
What is the salary range for this position?
The typical starting salary range for this role is €104.800 to €165.700 EUR in base salary, with no variable compensation component.
What core technologies will I be working with?
The codebase utilizes TypeScript and Python, and you will work extensively with Elasticsearch, public APIs, MCP tools, and various agent frameworks.
What kind of team environment should I expect?
You will work alongside data scientists, backend engineers, product managers, and UX designers in a distributed, async-first team that values a pragmatic, low-ego style.