Forward Deployed Engineer at Sierra: Role, Salary, and Interview (2026)
Most Forward Deployed Engineer articles start with the product. This one has to start with the invoice. Sierra sells its customer-facing AI agents primarily on outcome-based pricing — enterprises pay per successful resolution, not per seat or per API call. That single billing decision changes what the FDE job is. At a company that charges per seat, an engineer who improves an agent makes a customer happier. At Sierra, an engineer who improves an agent’s resolution rate directly moves the revenue line, because the agent only bills when it actually solves the customer’s problem. The FDE isn’t a post-sale cost center bolted onto a subscription. The FDE is the P&L.
That framing matters because Sierra, more than almost any company on the job board, is built around the deployment motion. Founded by Bret Taylor — chair of the OpenAI board, former co-CEO of Salesforce, former CTO of Meta, co-creator of Google Maps — and ex-Google VP Clay Bavor, Sierra launched publicly in early 2024 and crossed $100 million in annual recurring revenue about seven quarters later, a pace it describes as among the fastest in enterprise-software history. In May 2026 it raised a $950 million Series E led by GV and Tiger Global at a roughly $15 billion valuation, with over $1 billion in cash on hand. It reports around $200 million in ARR and counts more than 40% of the Fortune 50 as customers — SoftBank, Uber, Rivian, CLEAR, and Sutter Health among the named ones. This is not a lab renting out a model. It’s an applied-AI company whose entire business is agents that go live inside big enterprises and get paid for outcomes.
What the role actually is
Sierra doesn’t hire one flavor of deployment engineer; it hires a spectrum. At the customer-facing edge is the Agent Engineer — the person who designs, builds, and tunes the conversational agents that go into production for a specific enterprise. Alongside that sits the Agent Strategist, a hybrid of consulting, go-to-market, and hands-on agent building — closer to a solutions role than a pure IC one. And then there’s the variant currently on our board, the Forward Deployed Infrastructure Engineer, which owns the harder-to-fill infrastructure half of a deployment: private networking, scaling, latency, and reliability for agents running inside customer environments. The role is open in San Francisco, London, and Tokyo.
The distinction is worth sitting with, because it tells you how Sierra thinks about the work. Most FDE postings blend “talk to the customer” and “make it reliable” into one job description and hope the candidate is strong at both. Sierra has split off the infrastructure discipline into its own seat — which is the rarer skill set and, unsurprisingly, the one it pays a premium for. If your strength is production systems rather than stakeholder management, the infra variant is the one to target, and it’s less crowded than the agent-building track that every AI-curious full-stack engineer applies to.
One practical note: Sierra is a predominantly in-person company. The SF headquarters is the center of gravity, with growing offices in North America, Europe, and Asia. Expect this to be a badge-in role, not a remote one — a meaningful filter compared with the remote-friendly OpenAI and Cohere versions of the job.
Compensation
The cleanest number comes straight from the live posting, not an aggregator: the Forward Deployed Infrastructure Engineer role lists $230K–$390K base band in San Francisco, with London at roughly $170K–$290K and a Tokyo seat also open. That’s the infra variant; the customer-facing agent roles sit in a similar neighborhood before equity.
| Component | Figure (2026) |
|---|---|
| FD Infrastructure Engineer band (SF, from posting) | $230K–$390K |
| FD Infrastructure Engineer band (London) | ~$170K–$290K |
| Sierra median total comp (Levels.fyi aggregator) | ~$227K |
| Sierra SWE range, SF Bay Area (Levels.fyi aggregator) | ~$200K–$520K+ |
| Equity | Private, illiquid options at ~$15B valuation |
Treat the aggregator figures as directional and small-sample — Levels.fyi’s Sierra data reflects a young company with relatively few self-reported packages, so the top of the range is skewed by a handful of senior hires and shouldn’t be read as a typical FDE number. The posting band is the honest anchor. Where the real upside lives is the equity: options in a company that has roughly 150x’d its implied worth against early rounds and is compounding ARR fast. That’s the classic applied-AI trade — cash comp that’s strong but not frontier-lab-silly, paired with equity that could matter a great deal if the enterprise-agent thesis holds. As always, the move is to ask the recruiter directly about the current preferred-share price and whether refresh grants exist, because a $15 billion mark means your strike is no longer cheap.
The interview loop
Candidate reports (aggregated from Glassdoor, Exponent, and interview-prep write-ups) describe a loop that runs roughly two to five weeks and is refreshingly job-shaped. Read it as a pattern, not a guarantee.
| Stage | Format | What’s reportedly tested |
|---|---|---|
| Recruiter screen | 30 min | Background, role fit, in-person expectations |
| Technical screen | ~1 hr, live coding | A LeetCode-style problem — one report describes a cycle-detection problem “dressed up” as an Excel question |
| Take-home | Async | Build a small customer-support agent; reason about prioritization, metrics, and observability |
| Onsite | 3 × 1 hr | Agent design, production judgment, past experience, collaboration |
The take-home is the tell. Rather than a generic algorithms gauntlet, Sierra asks you to build a real (if small) support agent and then defend your choices about what to prioritize, how to measure success, and how you’d observe the thing in production. That mirrors the actual job — and it rewards candidates who think about evaluation and reliability, not just whether the demo runs once. The broadly reported theme is that Sierra weights practical engineering and product judgment over trivia and brainteasers. If you’ve never shipped an agent and reasoned about its resolution rate, the take-home is where that gap shows.
Who this role is for
The Sierra FDE seat rewards a specific profile: an engineer who genuinely enjoys the last mile — the part where a promising agent has to survive contact with a real enterprise’s edge cases, latency budgets, and skeptical operations team. Because pricing is tied to outcomes, your work is legible in a way it rarely is at seat-priced companies: the resolution rate you drag from 70% to 85% is money. That’s motivating for some people and stressful for others, and it’s worth being honest with yourself about which. If you want the infrastructure track specifically, lean into distributed-systems reliability and private-networking experience in your application — that’s the scarcer half of the role and the reason the band runs where it does. And if in-person in San Francisco, London, or Tokyo is a dealbreaker, this isn’t your seat, however good the equity looks.
Sierra’s Forward Deployed Infrastructure Engineer role is live on our job board now — or get the weekly digest to track new FDE openings as they post.