Agentic AI Engineer

Deep Seas•San Diego, CA

8h•Remote

About The Position

We're a small fast-moving team building AI Agents to solve real-world cybersecurity challenges - things like threat detection, log analysis, and incident response automation. This isn't about flashy demos. We care deeply about correctness, reliability, and systems that actually perform under pressure. We're looking for a talented engineer with expertise taking complex systems from design all the way to production. The right candidate is excited to tackle difficult challenges using a blend of out-of-the-box thinking and proven software engineering best practices.

Requirements

Strong Python (or equivalent modern language) engineering skills.
Deep understanding of language performance and internals (eg. GIL, concurrency tradeoffs, memory constraints).
Proven experience building production-grade AI agents. Not just demos. Systems that run, recover, and scale.
Thorough understanding of LLMs and agentic systems. You understand how they actually work under the hood.
Strong context engineering skills. Able to curate, compress, and structure large datasets for effective LLM use.
Experience evaluating, testing, and monitoring agents and LLM interactions using the latest techniques or frameworks.
Experience handling large-scale high-volume data workflows.
Portfolio of shipped work (GitHub, case studies, or equivalent).
Bachelor of Computer Science/Engineering or above.
Must be a US Citizen.

Nice To Haves

Experience with cybersecurity tools (CrowdStrike, Splunk, Mandiant, etc.)
Familiarity with multiple agent frameworks (Claude Agent SDK, LangGraph, AutoGen, custom systems, etc.)
Experience evaluating tradeoffs between frameworks vs. custom orchestration.

Responsibilities

Design, develop, and productionize advanced agentic systems that handle large volumes of security event data, operate with resilience in noisy, real-world environments, and continuously evolve as new attack patterns and tools emerge.
Improve and harden existing agentic systems.
Work with LLMs in production environments, not just prototypes.
Engineer robust context pipelines for large-scale data ingestion and reasoning.
Optimize systems for performance, reliability, and failure handling.
Evaluate and integrate new agent frameworks, tools, and approaches.
Rapidly hypothesize -> build -> test -> validate -> iterate.