Applied AI Engineer, Internal Systems

Gallatin•El Segundo, CA

6d•$80,000 - $210,000

About The Position

Every hour spent chasing information or moving it between tools is an hour someone can't spend on the work they were hired to do. This role exists to remove that overhead. As the Applied AI Engineer, Internal Systems you'll own the engineering behind the AI tools Gallatin uses to run the company. You'll work alongside the Internal AI Operations Product Manager to shape solutions, with the technical design and implementation in your hands. You'll also own deployment and reliability. Your users will be your colleagues. Work together with them to understand the bottlenecks, connect information across internal tools, and build agents that carry tasks through to completion.

Requirements

Shipped software people use and maintained it after launch.
Can turn an unclear problem into a technical plan and ship a useful first version.
Strong programming ability in Python or TypeScript.
Experience building APIs.
Experience working with databases.
Experience building LLM applications or agents that use external tools and data.
Can explain how LLM applications/agents were tested and where they fail.
Enough experience across the stack to build a usable interface and deploy the service behind it.
Engineering judgment about reliability and security.
Knows when a model needs a human check and when a simpler implementation will work better.
Comfort working directly with nontechnical colleagues.
Can explain a tradeoff clearly and change approach when the workflow demands it.

Nice To Haves

Experience building internal tools.
Experience evaluating AI outputs.
Experience integrating business systems.

Responsibilities

Own the engineering behind the AI tools Gallatin uses to run the company.
Work alongside the Internal AI Operations Product Manager to shape solutions, with the technical design and implementation in your hands.
Own deployment and reliability of AI tools.
Work with colleagues to understand bottlenecks, connect information across internal tools, and build agents that carry tasks through to completion.
Turn workflow problems into technical requirements.
Identify where a process should change before automating it.
Build applications and integrations that connect models to company systems.
Use conventional code where the behavior needs to be predictable.
Turn useful experiments into software people can depend on, with evaluations, observability, and a way to recover when something fails.
Design access controls and human review into workflows that handle sensitive information or take consequential actions.
Instrument the systems so we can measure time saved and output quality.
Use failure data to improve the software and give the product manager evidence for what to improve or retire.
Help colleagues use and extend what you build.
Document the systems so another engineer can maintain them.
Work with the Internal AI Operations Product Manager to choose an initial workflow with a team that will use it.
Establish a baseline together.
Own the technical design and ship a working version with evaluations and observability, then improve it through regular use.
Ensure that by day 90, it is known whether the workflow saves time and meets the team's quality bar, with a plan to keep it working.