Forward Deployed Engineer at Fin (Intercom): Role, Salary, and Interview (2026)
Most FDE teams sell hours of engineering. Fin sells resolutions. The company formerly known as Intercom prices its AI customer agent at $0.99 per conversation the agent actually resolves. There is no seat fee on that line item. That one pricing choice explains almost everything about the Forward Deployed Engineer job there. Every point of resolution rate an FDE adds to a large account is revenue Fin bills that month. Every point it fails to add is revenue it doesn’t.
It also explains why the job looks less glamorous on paper than the Sierra or frontier-lab versions, and why that’s not necessarily a bad thing.
The company you’d be joining
The corporate story moved fast in 2026. Intercom renamed itself Fin in May, a signal that the AI agent, not the help desk, was now the business. On June 15, Salesforce signed a definitive agreement to buy Fin for roughly $3.6 billion. The deal closed on September 10, well ahead of the window Salesforce gave at signing. Fin now runs as a unit inside Salesforce AI Labs rather than being folded into Agentforce. It still sells to companies on rival help desks, and per-resolution pricing carried over unchanged at close.
The scale numbers are company claims, so read them that way. At acquisition, coverage put the whole business at more than $400 million in ARR and over 30,000 customers. The Fin agent product alone was approaching $100 million in ARR. Fin cites an average resolution rate in the high 60s to mid 70s, depending on the date and the source, up from roughly 25% at launch. Its own post-trained model, Fin Apex, is pitched as beating general-purpose frontier models on customer-service resolution benchmarks.
For a job seeker, the acquisition matters in two ways. First, the FDE team is now inside a company that is itself hiring Forward Deployed Engineers by the hundreds for Agentforce. Fin’s version is the narrower, more specialized seat. Second, the postings still list RSUs. Ask your recruiter directly whether those are now Salesforce RSUs and on what vesting schedule, because the answer changes how you value the offer.
What the role actually is
Fin’s FDE motion is about two years old. Diego Ballona, the Senior Director of Engineering who runs it, said on Gainsight’s [Un]Churned podcast that it began as a five-customer design-partner experiment. Two details from that conversation are worth more than any job description.
First, the work is mostly configuration, not custom code. FDEs tune prompts, run experiments on agent behavior, wire up data and actions, and teach the customer’s team to manage their own agent. When a pattern repeats across customers, it turns into product. An AI-scored CSAT feature and Fin for Slack both came out of FDE engagements.
Second, engagements are bounded on purpose. Ballona argued against FDEs becoming open-ended managed-service operators for a customer’s agent. The FDE gets the account to value early, then hands it back to customer success.
The team structure is unusual too. Fin works in small cross-functional units it calls “swarms,” one per large engagement. A June 2026 post on Fin’s own blog describes each swarm as a trio. The forward-deployed engineer configures Fin for the customer’s environment. A forward-deployed product manager runs the relationship and the workstreams. A forward-deployed data scientist defines the success metrics and proves whether they moved. Proof-of-concept engagements reportedly run about two weeks. The same post cites one account where resolution rate went from about 50% to over 65% within weeks of optimization.
That split is the contrarian point. At most companies the FDE is expected to be engineer, PM, and analyst at once. At Fin, measurement belongs to someone else, and the FDE’s job is narrower and more technical than the title suggests. If you like owning the whole customer story, that’s a constraint. If you’d rather go deep on making an agent behave, it’s a feature.
There’s also a pre-sale flavor. A Senior Manager posting from October 2026 leads an “FDE New Business” team focused on large, complex deals. That’s a player-coach seat working alongside Sales and Solutions, so some FDEs are in the room before the contract is signed.
Compensation
Fin publishes base ranges on its US postings. They are lower than the AI-lab headline numbers, and the gap is worth naming.
| Role (posting) | Location | Base range |
|---|---|---|
| Forward Deployed Engineer | San Francisco Bay Area | $180,000–$237,000 |
| Senior Forward Deployed Engineer | San Francisco Bay Area | $215,000–$237,000 |
| Senior Forward Deployed Engineer | Chicago | $190,000–$215,000 |
| Senior Manager, FDE (New Business) | San Francisco Bay Area | $252,000–$280,000 |
| Senior Forward Deployed Engineer | Dublin | Not published |
On top of base, the postings mention eligibility for a corporate bonus or sales incentive, plus RSUs. A $237K ceiling for a senior FDE in San Francisco sits well below Sierra’s $230K–$390K infrastructure band and below the frontier-lab ranges in our salary guide. In exchange, you get a public-company employer and far fewer layoffs-by-runway risks than at a Series B.
Dublin is the open question. Fin doesn’t publish Irish bands. Levels.fyi’s self-reported data for Intercom software engineers in Ireland shows a median total comp of about €115K, with senior product engineers around €126K. Treat that as a directional floor, not an FDE number: it’s a small sample, it’s for product engineers, and it predates the acquisition. It’s still one of the stronger FDE options with a real European base. Our Europe guide has the rest of the field.
The interview loop
Here we have to be honest: there are no reliable public reports of Fin’s FDE-specific loop yet, and we won’t invent one. The postings themselves say the role requires:
- deep knowledge of one high-level language (JavaScript, Python, or Ruby)
- fluency with REST APIs
- the ability to prototype and ship quickly without dropping quality
- the ability to explain technical trade-offs to non-technical stakeholders
- willingness to travel to customer sites
The senior Dublin posting adds a preference for former founders and early-startup engineers.
Read those against the swarm model and the preparation writes itself. Expect to be tested on whether you can take a vague customer complaint (“the bot keeps escalating refund questions”), find the cause in prompts, knowledge content, or a broken action, and fix it fast. Practice debugging an agent’s conversation logs, not inverting binary trees. And prepare one story where you moved a business metric, not just shipped a feature. Our interview guide covers the general FDE format.
Who this role is for
Fin’s FDE seat suits an engineer who likes tight feedback loops and doesn’t need the title to mean “does everything.” You’ll spend your days making a production agent resolve more of a real company’s support volume, with a data scientist keeping score and a PM keeping the customer happy. The pay is solid rather than spectacular. The product is one of the most widely deployed AI agents anywhere. And because every resolution is billed, it’s one of the few places where you can see what your work is worth in plain numbers. If you want frontier-lab comp or unscoped ownership, look elsewhere. If you want to get very good at the part of AI deployment that actually pays, this is a strong place to learn it.
Fin’s Forward Deployed Engineer roles are on our job board, or get the weekly digest to catch new openings as they post.
By A. George. Postings read November 2026 on Fin’s Greenhouse board and job aggregators. Fin revises requisitions often and some listed here have closed, so confirm bands on the live listing.