Forward Deployed Engineer at Perplexity: Role, Salary, and Interview (2026)
Perplexity doesn’t have a Forward Deployed Engineer job. It has three of them, and they don’t do the same work. One sits on the Sonar API team, diving into logs to fix retrieval latency for developers. One carries the internal title “Member of Technical Staff” and embeds with enterprise customers wiring Perplexity’s search into internal knowledge workflows — a New York seat, the company’s first FDE hire in the city. The third calls itself an “Applied AI Architect” and belongs entirely to Computer, the agentic product Perplexity shipped in February 2026 that’s since become the company’s biggest growth driver. Three roles, three product surfaces, one job family assembled on the fly. That’s not sloppy naming — it’s what it looks like when a consumer search company backs into an enterprise deployment function faster than it can standardize the title.
What the role actually is — all three versions
The Sonar posting is the most conventional: officially a “Customer Success Engineer,” internally understood as a forward-deployed seat, it has you acting as technical partner to developers building on Perplexity’s search API. The job is API integration, auth flows, and latency debugging — “diving directly into code and logs” per the listing — plus feeding customer pain back to the API team. It’s the closest thing here to a classic solutions-engineer job, just pointed at a fast-moving developer product.
The Applied AI MTS role, based in New York, is the one that actually matches the “forward deployed engineer” label in the market’s usual sense. Per the posting (since closed to applications) mirrored here, it’s on-site, requires 5+ years across software, forward-deployed, or solutions engineering, and has you building “end-to-end connections between Perplexity’s platforms and customer systems” — including, notably, workflow orchestration for Perplexity Computer’s agentic capabilities inside enterprise environments. Wrapping the FDE function in an MTS title is a lab-branding move borrowed from research labs like OpenAI and Anthropic, applied to a job that is, functionally, enterprise deployment.
The Computer Applied AI Architect seat, based in San Francisco, is the newest and most senior-leaning of the three. It skips the coding-heavy framing entirely in favor of “technical discovery,” designing agentic workflow solutions for strategic accounts, and converting bespoke implementations into reusable product assets — closer to Glean’s “founding forward deployed engineer” than to a delivery role. Given that Computer only launched in February 2026, this seat is effectively defining what enterprise Computer deployment looks like as it goes.
The throughline: Perplexity is a 100-million-monthly-user consumer product that has, in under a year, grown a full forward-deployed org chart to sell an agent product to enterprises. The function is younger here than at Palantir or even Anthropic, and it shows — three job titles doing three adjacent jobs is what an improvised motion looks like before it consolidates.
Compensation
None of the three postings disclose a salary band directly. The most useful outside signal comes from Aced’s (formerly Exponent) analysis of 135 US forward-deployed-engineer postings in mid-2026, which put Perplexity’s posted base band at the very top of the market:
| Company | FDE base salary range |
|---|---|
| Perplexity | $220,000–$405,000 |
| Anthropic | $280,000–$320,000 |
| $174,000–$364,000 | |
| OpenAI | $153,000–$325,000 |
| Sierra | $180,000–$390,000 |
| Market median (all companies) | ~$190,000 (25th–75th pct: $160K–$220K) |
Treat the top of that Perplexity range with some skepticism — it’s an aggregator’s read across postings, not a single disclosed band, and it likely reflects the spread across all three roles above rather than one job at one level. General engineering compensation at the company, per aggregator estimates from Glassdoor (self-reported base only) and compensation trackers, spans roughly $160K–$270K base for individual contributors up to $250K–$350K at staff/principal, with total comp reportedly running $350K–$550K once equity is priced in — though these are third-party estimates, not company disclosures, so read them as directional.
The equity conversation is where Perplexity’s story gets more interesting than most FDE seats. The company raised $200 million in September 2025 at a $20 billion valuation; multiple 2026 trackers now put the mark closer to $22–23 billion. Revenue is the reason: annualized revenue reportedly hit roughly $750 million by August 2026, up from about $232 million at the end of 2025 — a jump growth trackers attribute directly to the February 2026 launch of Computer and a new usage-based pricing layer. That’s a company whose enterprise agent product is less than a year old and already reshaping its revenue mix, which is exactly the kind of inflection point that makes FDE equity worth more than the base salary line suggests.
The interview loop
Perplexity doesn’t publish an FDE-specific process for any of the three roles, so what follows is the company’s general engineering loop, reported by candidates — not a confirmed FDE script, and the two consultative roles (Applied AI Architect, Sonar) likely add a discovery or stakeholder-facing round the standard loop doesn’t cover.
| Stage | Format | What’s tested |
|---|---|---|
| Recruiter screen | 20–30 min | Product knowledge — expect to be asked what you’d fix about Perplexity |
| Technical phone screen | 45–60 min | One applied Python coding problem; production quality over algorithmic tricks |
| Onsite coding (x2) | 60 min each | Applied scenarios — retrieval pipelines, tokenizer logic, state machines |
| ML/search systems deep-dive | 60 min | BM25 vs. dense retrieval, ANN indexing, hybrid search, reranking, eval metrics |
| Product & craft | 45–60 min | Critical product evaluation — “where does Perplexity lose to Google today?” |
| Founder/leadership round | 30–45 min | Technical depth and product conviction, often with CEO Aravind Srinivas or a founding engineer |
Candidates report the whole loop compresses into four to six rounds over one to three weeks — fast for a company at this valuation, and a signal that Perplexity is still optimizing for speed of hiring over process maturity. The ML/search deep-dive is the highest-filtering round for engineering hires generally; for the two customer-facing seats, expect that technical bar paired with a much harder read on whether you can translate a retrieval-quality problem into language an enterprise buyer’s stakeholders will actually act on.
Should you take it?
The case for: Perplexity’s posted FDE band tops the market, the equity sits inside a company whose enterprise revenue is compounding off a four-month-old product, and having three distinct role flavors means you can pick the surface — API support, enterprise integration, or ground-floor agent deployment — that matches how you actually want to spend your day. The Computer Applied AI Architect seat in particular is about as close to “define the playbook” as this job gets anywhere outside Glean.
The case for caution: none of the three postings disclose comp, two require on-site presence in New York or San Francisco, and a function three job titles deep into improvisation is a function without a settled career ladder yet. If you want the structured, mature version of this job, Anthropic and Palantir have had longer to build one. If you want to help build the ladder itself at a company growing revenue 3x in eight months, this is one of the more interesting bets in the market right now.
Tracking FDE openings at Perplexity and 80+ other companies — see the live job board or get the weekly digest.
— A. George