Home / Frontier practices / Agentic AI & AI Systems
AGENTS / PRODUCTION INTELLIGENCE

Agents that do
real work.

Beyond prototypes, production agents demand deep engineering across tools, memory, orchestration, evaluations and security. We recruit the researchers and software engineers who can deliver dependable autonomous workflows.

Impax
01 / THE TECHNICAL LANDSCAPE

Specialism is
the difference.

Search quality depends on understanding what the team is actually building. We map the technical problem first, then identify the skills, research evidence and operating style that matter most for the hire.

Our scope spans the following areas, with each brief shaped around the specific stage, stack and constraints of the business.

  • 01Agent architectures and orchestration
  • 02Evaluation harnesses and observability
  • 03Tool use, MCP and integrations
  • 04Memory, retrieval and context engineering
  • 05Reliability, latency and safety
  • 06Human-in-the-loop product workflows
02 / REPRESENTATIVE MANDATES

Roles with
real consequence.

These are examples of role types within our specialist coverage, not claims that each role is currently open. For active searches, visit the opportunities board.

  • Founding AI Agent Engineer
  • Staff Agent Infrastructure Engineer
  • AI Evaluation Engineer
  • Applied AI Research Engineer
  • Forward Deployed AI Engineer
03 / HOW WE ASSESS

Signals over
superlatives.

Our approach starts with the evidence a hiring manager would actually want to see. These are sample questions that help us evaluate depth; they are not universal pass/fail rules.

Which workflows are actually live with users?+

A production agent should have actual users or workflows, measurable outcomes and a clear failure-handling strategy. We distinguish these from narrow demo workflows.

How are failure modes observed, replayed and improved?+

Replayable traces, scenario sets, offline and online evaluations, human review and production monitoring can each be relevant. We ask which choices improved reliability.

Where do they draw the line between a deterministic system and an agent?+

The right system may involve deterministic software, model-led decisions or a combination. The strongest candidates can explain why each element is present.

04 / CONNECTED DISCIPLINES
BUILDING IN THIS SPACE?

Let’s discuss the people
who will move it forward.

Speak with Impax ↗
A CLOSER LOOK / ROLE CALIBRATION

From agent prototype to production system

Agentic systems involve more than connecting an LLM to a tool. Teams must decide how tasks are decomposed, tools are selected, state is maintained and failure is detected. The best engineer for a customer-facing workflow may differ from one optimising evaluation infrastructure or building an internal developer platform.

For production mandates we look for evidence of traceability, reproducible evaluation, retry and recovery logic, memory design, orchestration choices, latency control and a clear view of model limitations. We also ask what the engineer owned rather than assuming adoption of an agent framework equals production expertise.

REPRESENTATIVE MANDATES

The talent we help identify.

01Agent systems & applied AI engineers
02Agent evaluation / reliability specialists
03LLM platform and orchestration engineers
04Product engineers for AI-native workflows