Agent Infrastructure Engineer

Nebula; El Segundo (CA); Hybrid; USD 100,000 - 250,000; Full time

Benefits offered by this job

Equity participation in a high-growth 

Comprehensive health, dental, and VIs

401(k) with company matching

Bonus for living within 5 miles of El 

On-site work with rare work-from-home

Merit-based organization where your‑to

Job summary

Nebula in El Segundo is building Merovingian, the AI side of our internal platform. You will design and implement the runtime, identities, and permissions that let agents work inside the company as colleagues of people here, and craft the trust model that keeps them secure.

The work spans distributed systems and security: identities, leases, policy-driven access, audit trails, and the evidence layer behind every call. You will own the execution plane and shape how agents and people share work.

Qualifications

• Bachelor's degree or equivalent experience in computer science or related field.
• Hands-on production distributed systems experience (leases, expiry, idempotency, back-pressure).
• Security and access-control engineering experience (authentication/authorization, audits).
• Comfort with Rust or another systems language (parsers, interpreters, state machines).
• Strong observability skills for telemetry, tracing, and cost accounting.

Responsibilities

• Develop the agent runtime, including parser/interpreter and execution loops.
• Define identity/lease models, heartbeats, and exactly-once delivery.
• Implement permission and trust models with need-to-know scoping and sandboxing.
• Ensure verification, safe handover, and audit trails for actions.
• Build telemetry, event taxonomy, tracing, and evaluations.
• Manage session context, retention, and cross-session notes.
• Define tool interfaces and access-control schemas.
• Design organization structure for multiple agents with proper supervision.

Skills

Distributed systems

Security engineering

Access control

Rust

Observability

Threat modeling

Adversarial testing

Education

CS or related degree, or equivalent experience

Job description

What to Expect

Your role is to build the runtime, the identities and the permissions that let agents work inside the company as colleagues of the people here, and the trust model that keeps them inside it.

Merovingian is the AI side of our internal platform: agents run as emulators with their own identities. Instead of expecting agents to carry the process inside them, we wrote the process down outside them, as code, and made the agent an executor inside it.

The work is distributed systems and security: identities and leases, delivery and supervision, permissions enforced as policy rather than instruction, and the evidence layer behind every call. You will own the execution plane and the design of how agents and people share work.

We are looking for engineers with experience and interest in distributed systems, security and access control, the evaluation and observability of agent systems, runtime or language implementation, and how an organization should be structured.

Every role works with our AI systems daily. What can be deterministic, must be. You need no AI background; we prefer people without one. We hire for your knowledge and experience in the field first, so you can steer the ship; the tooling is a learning curve we expect you to take on.

We work on-site in El Segundo. This is hard work, but you will be rewarded with equity in a company we believe will become one of the world's most valuable.

What You'll Do

• The agent runtime: parser and interpreter, execution loops, context compaction, provider routing, inter-process channels
• Identity and the collective: allocated identities, roles as leases, heartbeats and takeover, exactly-once delivery
• The permission and trust model: need-to-know scoping, per-tool checks, escalation flows, sandboxing, audit trails
• Verification and safety: independent confirmation a write landed, refusal paths, handover to a person
• Telemetry and evaluation: event taxonomy, tracing, cost accounting, transcript review, rubrics and offline evaluations
• Memory: session context and its lifecycle, notes carried between sessions, retention and visibility rules
• Tool interfaces: schemas for context economy, errors an agent can act on, access control
• Agent organization design: roles and boundaries, escalation and handoff, supervision of many agents

What You'll Bring

• A degree in computer science, computer engineering or a related field, or equivalent experience
• Hands-on production distributed systems: leases and expiry, idempotency, back-pressure, process supervision and their failure modes
• Security and access-control engineering: authentication and authorization, capability scoping, secret handling, sandboxing, audit trails
• Comfortable close to a runtime in Rust or another systems language: parsers, interpreters, state machines
• Fluent with observability of systems whose output is text: structured events, tracing, cost accounting, evaluations
• The judgment to decide what a person confirms, what an agent may do alone, and why
• Bonus: threat modeling a privileged automated actor, formal specification, or adversarial testing of language models

Expected Compensation

$100k to $250k base salary, plus equity

• Equity participation in a high-growth startup
• Comprehensive health, dental, and vision insurance
• 401(k) with company matching
• Bonus for living within 5 miles of our El Segundo facility
• On-site work with rare work-from-home exceptions
• Merit-based organization where contribution drives reward