Modal
1 open Forward Deployed Engineer role at Modal. Curated with comp bands.
Comp on the board: $180K–$250K + equity.
On the board since Jul 2026 · still hiring
Our take
Why Modal
The serverless GPU cloud behind the AI-coding boom — where 'forward deployed' means inference-optimization work alongside frontier labs, with the team shipping open source.
- FDE type
- Solutions engineering
- Travel
- Medium
- Comp
- Competitive
- Stage
- Growth-stage
A good fit if you're…
- Systems engineers who want real GPU, inference, and training-performance work
- People who'd rather contribute to open source than produce customer decks
- Engineers who want frontier-lab and hot-startup customers without joining a lab
- Candidates open to strictly in-person work in NYC, SF, or Stockholm
The serverless GPU cloud under the coding boom
Modal is a serverless compute platform for AI workloads: developers run inference and training on its GPUs, and it also serves the sandboxes where AI-generated code executes. Modal doesn't own hardware — it rents from 13 infrastructure partners, up from five. CEO Erik Bernhardsson (ex-Spotify ML) has said the last six months of enterprise AI-coding adoption 'have been driving everything,' and the FDE seat sits directly on that demand.
Forward deployed means GPU-level performance work
This is unusually deep FDE work. Rather than integrations and slideware, you tune production AI workloads at the systems level — inference optimization, model serving, GPU programming, and training infrastructure — for customers pushing real scale. The deliverable is a faster, cheaper, more reliable deployment, not a merged ticket. It's closer to a performance-engineering embed than a classic solutions role.
Who you'd actually work with
The posting names the accounts directly: you architect and optimize production AI workloads hands-on with companies like Suno, Lovable, Cognition, Ramp, and DoorDash. That's a front-row seat to how the best AI products run in production — the kind of customer roster most solutions engineers never touch — and it's why the role rewards people who want frontier problems without the frontier-lab headcount politics.
Systems chops over ML research
Modal asks for only ~2+ years of professional ML engineering, ideally in inference optimization, training, GPU programming, or ML infra — so it favors strong systems engineers over PhD researchers. Familiarity with serving stacks (vLLM, SGLang) and training frameworks (slime, verl, TRL) is expected, and the team are active SGLang contributors. Published work and open-source contributions earn a bonus.
Comp and where you'll sit
The board shows $180K–$250K plus equity — mid-pack cash for the tier, but the equity is the real story given the trajectory below. Work is in-person in New York, San Francisco, or Stockholm; the Stockholm option makes this one of the few FDE seats on the board with a European base. Expect medium travel, weighted toward customer sites and deployment sprints.
The trajectory you're buying equity into
Modal's annualized revenue jumped roughly 5x — from ~$60M in September 2025 to ~$300M by May 2026 — as its valuation climbed from $1.1B to $4.65B in eight months. By late September 2026 it was reported to be in talks for a new round near a $15–16B valuation. Few AI-infra companies are compounding this fast, which is what makes the equity portion worth weighing seriously.
Questions candidates ask
How much ML experience do I really need?
The posting asks for ~2+ years of professional ML engineering, ideally in inference optimization, model training, GPU programming, or ML infrastructure. It's aimed at strong systems engineers, not research scientists.
Is this a real FDE role or a performance-engineering job?
Both. It's forward deployed in that you work directly with named customers like Suno, Lovable, Cognition, Ramp, and DoorDash — but the work itself is GPU-level performance and serving optimization, not integrations or decks.
Where can I be based, and is it remote?
It's in-person in New York, San Francisco, or Stockholm. The Stockholm option is unusual — one of the few European-based FDE seats on the board.
Do FDEs here work on open source?
Yes. Team members are active contributors to SGLang, and the role explicitly rewards side projects, open-source contributions, and published work.
Analysis updated 2026-09-29
Open roles at Modal
Compare Modal
Head-to-head
Others in the same tier
Free forever. Unsubscribe anytime.