Forward deployed engineer at Decagon
At Decagon, the forward deployed engineer role is the Agent Deployment Engineer (ADE): a technical, customer-facing engineer who owns the build of an enterprise customer's AI agent, from scoping and integrations through evaluation, launch and tuning, working alongside an Agent Product Manager.
Checked
Decagon, a company that builds AI agents for customer service, split its deployment work into two roles. In an August 2026 post on Decagon's blog, Alex Lewis, Manager of Agent Deployment Engineering, explains that Agent Product Managers (APMs) found the customer relationship and the agent build each needed sustained depth, so Decagon created Agent Deployment Engineering to own the technical build.
In that split, APMs lead the customer relationship and the wider transformation; ADEs turn the customer's context into a production-ready agent on Decagon's platform.
The titles Decagon uses
Agent Deployment Engineer (ADE)
Posted on 1 November 2026 in San Francisco, New York, Toronto, London and Australia, with German-, French- and Spanish-speaking versions.
Agent Product Manager (APM)
The customer-facing partner role, posted at Agent Product Manager and Senior Agent Product Manager levels.
Related roles
Agent Development Manager, Agent Strategy Manager and Engineering Lead, Strategic Deployments also appear on Decagon's board.
What the job involves
The agent build, end to end
ADEs own the architecture and execution of agent builds for enterprise customers, from systems design and scoping through implementation, evaluation and production deployment.
Agent logic and guardrails
The post covers designing the agentic logic that governs behaviour, and layered guardrails and supervisory controls for safe, compliant and predictable performance.
Integrations
ADEs build and test integrations with customer systems such as CRMs, ticketing systems and data pipelines, including the tools and APIs the agent needs.
Evaluation and debugging
ADEs design evaluation and regression tests against ground truth, test agents against real conversations, and trace failures through execution logs and model behaviour.
Patterns back to the product
According to the blog, recurring lessons become reusable architecture patterns, tools and testing practices, and recurring gaps go to Decagon's product and engineering teams.
Pay, requirements and openings
Posted pay (Ashby, read 1 November 2026)
$175,000 to $230,000 plus equity in San Francisco and New York; £110,000 to £150,000 in London; CA$150,000 to CA$200,000 in Toronto; A$213,000 to A$266,000 in Australia.
Requirements
Five or more years in a technical customer-facing role such as solutions engineering, forward deployed engineering, technical consulting or implementation engineering, with comfort writing code and building integrations. LLM and agent experience is listed as a plus. Decagon describes itself as an in-office company.
Interview process
Decagon's posts and blog do not describe the interview loop.
Where to find openings
Decagon's careers page at decagon.ai/careers lists open roles from its Ashby board at jobs.ashbyhq.com/decagon, filtered by team. Both were live on 1 November 2026.
How it compares with FDE roles elsewhere
The ADE is a forward deployed engineer with a narrower brief: the technical build of one kind of system, a customer-service agent, on Decagon's platform. The customer relationship that a Palantir Delta or an OpenAI FDE often carries alone sits with the APM instead.
Sierra's Software Engineer, Agent is the closest equivalent. Decagon's posted band of $175,000 to $230,000 is narrower than Sierra's $180,000 to $390,000, and the ADE asks for five years of customer-facing technical work, where Sierra also hires new grads.
For companies that want FDE capacity without hiring
Decagon's ADEs build agents on Decagon's platform for Decagon's customers. A company that wants an agent built inside its own product, with the evals and integrations staying in its own codebase, has another route.
Motiontic embeds founder-level engineers in a client's codebase and team chat on a monthly engagement. The work is built on whichever models and cloud fit the problem, and the documentation, tests and evals stay with the client's own team. The model suits a defined build that has to ship more than open-ended headcount.
Agent Deployment Engineer at Decagon compared with the FDE role generally (checked 1 November 2026)
| Decagon | FDE role generally | |
|---|---|---|
| Title | Agent Deployment Engineer (ADE) | At least six titles for the same job, per Perspective AI (June 2026) |
| What gets built | Customer-service AI agents on Decagon's platform | A production system for one customer |
| Who owns the customer relationship | The Agent Product Manager | Often the FDE, sometimes with a strategist |
| Core work | Agent logic, guardrails, integrations, evals, tuning after launch | Integrations and a production system through go-live |
| Experience asked | Five or more years in a technical customer-facing role | From new grad (Palantir) to five or more years (OpenAI) |
| US posted pay | $175,000 to $230,000 plus equity | Levels.fyi median total pay $220,000 across 163 salaries |
| Time split | Not stated | About 60% customer-facing, 30% deployment code, 10% internal (Perspective AI) |
Common questions
What is an Agent Deployment Engineer at Decagon?
Decagon's forward deployed engineer: a technical, customer-facing engineer who owns the build of an enterprise customer's AI agent, including its logic, guardrails, integrations and evaluations, alongside an Agent Product Manager.
How much does a Decagon Agent Deployment Engineer make?
Decagon's San Francisco and New York ADE posts listed $175,000 to $230,000 plus equity on 1 November 2026, and its London post listed £110,000 to £150,000.
What is the difference between an ADE and an APM at Decagon?
According to Decagon's blog, the Agent Product Manager leads the customer relationship and the broader transformation, while the Agent Deployment Engineer turns that context into a production-ready agent.
Related
Forward Deployed Engineer
Embedded in the client's codebase, Slack and standups, shipping against the client's own priorities.
Forward deployed engineer salary
FDE pay by source, company, stage and city, and what a hire costs compared with embedding.
FDE at Sierra
Sierra's Agent Engineers build and tune customer-facing AI agents with each customer.
FDE at Ramp
Ramp's FDEs ship product for its largest customers and build agents that do the work.
FDE at Anthropic
Applied AI engineers who build on Claude inside a customer's own systems.
Sources
- Decagon blog: Getting closer to the build, Alex Lewis (2026-08-12)
- Decagon (Ashby): Agent Deployment Engineer, San Francisco
- Decagon (Ashby): Agent Deployment Engineer, London
- Decagon careers
- Sierra (Ashby): Software Engineer, Agent, San Francisco
- Levels.fyi: Forward Deployed Engineer pay (US)
- Perspective AI: 2026 FDE hiring trends (2026-06-08)
Running agents that have to hold up?
Describe the agent, the systems it touches and what it must never get wrong. The first call is with a lab engineer, who will say plainly whether it is ready for production, and whether Motiontic is the right team.