Forward Deployed Engineer (Systems)

REALM; New York (NY); On-site; USD 180,000 - 240,000; Full time

Job summary

REALM in New York is seeking a software engineer to bridge research and customers, turning new agent methods into production-grade code inside client environments. You will design workflows, build integrations, and ensure reliability at scale across large data volumes.

You will work close with research and engineering, ship features to Fortune 100 customers, and own end-to-end delivery from prototype to live deployment. On-site in Manhattan with relocation supported for the right candidate.

Qualifications

• Three years or more writing software professionally.
• High-level Python skills including concurrency and data manipulation.
• Experience with distributed systems and large-scale pipelines.
• Proven ability to deliver a project to customers.

Responsibilities

• Write production code inside a customer environment and build agent workflows.
• Design and tune deployment workflows and pipelines.
• Integrate systems to support large data volumes across machines.
• Diagnose failures across layers and communicate with customers' senior engineers.

Skills

Python

Distributed systems

Data manipulation

Tools

Kubernetes

Terraform

CI/CD pipelines

Job description

REALM are partnered with a venture backed AI company in New York, building agents that work over operational data inside very large enterprises. Small team, research heavy, with engineers drawn from infrastructure and data companies you would recognise. They are well past the pilot stage. The product is live and growing inside Fortune 100 engineering organisations, and most of that growth has come from customers choosing to use more of it rather than from a sales team pushing it. I am glad to go through the detail on a call.

About the Role

You would sit between the research team and the customer. Research produces new approaches to how the agent reasons, evaluates and decides. Your job is to make those approaches survive contact with a real environment, where the data is raw, access is restricted, and the volume is enormous. That means writing production code inside a customer's own systems. Designing the workflow the agent follows and tuning it until the output is reliable. Building whatever integration or pipeline the deployment needs. Working at whichever layer the problem happens to be, rather than the one you are most comfortable in. It is a building job with customers present, not a pre-sales or solutions role. What this team discovers in the field determines what everyone else in engineering builds next quarter.

Responsibilities

• Three years or more writing software professionally. Two if the rest of the picture is exceptional.
• Python at a genuinely high level, including concurrent code and heavy data manipulation.
• Systems experience: services running across many machines, or pipelines moving very large volumes. Highly experienced in distributed systems.
• Applied work with language models. Designing prompts, wiring up tool calls, building agent workflows. Research credentials are not needed and not expected.
• Something you designed yourself and took all the way to customers relying on it.
• The ability to locate a failure without knowing in advance which layer it lives in.
• Enough confidence to hold a technical conversation with a customer's senior engineers, and enough patience for a situation that is rarely fully defined.

Qualifications

• Helpful, not necessary: Time spent close to production monitoring or telemetry systems.
• Container orchestration or infrastructure as code.
• Open source work on agent tooling.

Required Skills

• Five days a week on site in Manhattan. Relocation supported for the right person. Visa transfers have been done before.
• The interview is a three part simulation of the job, worked through with one of their engineers. Real problems, increasing in difficulty. No algorithm puzzles.

Equal Opportunity Statement

We are committed to diversity and inclusivity in our hiring practices.