Asks for

The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Comfortable with containers, Kubernetes, and cloud networking and security fundamentals”
The ad says“Comfortable with containers, Kubernetes, and cloud networking and security fundamentals”
The ad says“Comfortable with containers, Kubernetes, and cloud networking and security fundamentals”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Practical familiarity with the open-model ecosystem: Hugging Face, the major open weight families, and how teams evaluate one model against another”
The ad says“Comfortable with containers, Kubernetes, and cloud networking and security fundamentals”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Genuinely hands-on: you write Python comfortably and build your own demos rather than requesting them”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”
The ad says“Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost”

About Featherless

Featherless.ai is building the world's most reliable open-model inference platform.

Backed by AMD, Airbus Ventures, 500 Global, Kickstart Ventures, HF0 Residency, Panache Ventures, and Oakseed Ventures, we're a well-funded team of researchers and engineers on a mission to democratize AI through performance and accessibility.

Our cloud provides instant access to 40,000+ open-source AI models and the research innovations powering next-generation efficiency, reliability, and model optimization.

About the Role

We're hiring our first Sales Engineer to pair directly with our founding Account Executive in North America. You are the technical half of a two-person deal team: the AE owns the commercial motion, you own the technical win.

Our buyers are developers, ML engineers, and CTOs, and they don't buy on slides. They buy when someone credible sits with them, looks at their workload, and shows them that open models on Featherless are faster, cheaper, and more reliable than what they're running today. That's your job — hands on keyboard, in the customer's stack, from first technical call through production rollout.

Because you're the first SE here, you'll build the function as you go: the demo environment, the POC playbook, the benchmark harness, the security answers, the migration guides. Every deal you win should make the next one easier to win.

If you like being the person who can actually answer the hard question in the room — and want the technical foundation of a Series A GTM engine to be yours — this is for you.

What You'll Own Immediately

Partner with the founding AE as the named technical owner on every active opportunity in North America
Run technical discovery: map each prospect's models, workloads, latency and throughput targets, spend, and constraints
Design and drive POCs and benchmarks that prove Featherless against the incumbent — closed-model APIs, another inference provider, or self-managed GPUs
Build the demo environment and reusable technical collateral that the whole GTM team sells with
Own technical objection handling end-to-end: performance, reliability, cost modeling, security, and data handling
Be the field's voice into product and engineering — you'll know what we're losing on before anyone else does

Core Responsibilities

Lead technical calls tailored to each buyer's stack: architecture reviews, live demos, and working code against their real use case
Build migration paths off closed-model APIs and onto open weights — model selection, evaluation, prompt and output parity, cutover plan
Run benchmarks and produce the throughput, latency, and cost-per-token analysis that anchors the business case the AE builds
Scope and execute POCs with clear technical success criteria, then hold the customer and us to them
Write the technical sections of proposals, RFP responses, and security questionnaires; keep a reusable answer library so we never write the same answer twice
Support onboarding and first production workloads, then hand off cleanly and stay available for expansion
Feed structured field input to product and engineering — feature gaps, model coverage requests, reliability issues, competitive intel
Build and maintain the technical assets that scale the team: demo apps, notebooks, reference architectures, integration guides, internal enablement for BDRs and future AEs
Represent Featherless technically at conferences, meetups, and developer events
Keep POC and technical-stage detail current in HubSpot so forecasting reflects technical reality, not optimism

What You Bring

3–6 years in pre-sales engineering, solutions architecture, or a forward-deployed / customer-facing engineering role — at a GPU cloud, inference provider, MLOps platform, AI/developer tooling company, or cloud infrastructure vendor
Genuinely hands-on: you write Python comfortably and build your own demos rather than requesting them
Working knowledge of modern LLM inference — serving stacks (vLLM, SGLang, TensorRT-LLM or similar), OpenAI-compatible APIs, quantization, LoRA and fine-tune serving, batching and KV cache behavior, and what actually drives tokens/sec and cost
Practical familiarity with the open-model ecosystem: Hugging Face, the major open weight families, and how teams evaluate one model against another
Comfortable with containers, Kubernetes, and cloud networking and security fundamentals
Credible with both audiences in the same meeting — the ML engineer who wants the numbers and the CTO who wants the risk and cost story
Strong written communication: your benchmark writeups and architecture docs should be good enough to forward to a buyer's CEO
Entrepreneurial and self-directed — comfortable being the first SE, with no playbook and no one to escalate the technical answer to
Use AI tools heavily in your own workflow to research, prototype, and move faster

Nice to have: experience selling or building on AMD GPUs / ROCm; exposure to enterprise security and compliance review; open-source contributions or public technical writing; experience as the first technical hire on a GTM team.

Why Featherless

Build the sales engineering function at a company in one of the fastest-moving spaces in AI
Work as a true pair with the founding AE, and directly with the CRO and founders — small team, no layers, immediate impact
Real technical depth: 40,000+ open models and an in-house research team shipping inference and optimization work you'll get to sell
Be part of a small, high-performing team making open weight AI accessible to everyone
Competitive base + variable tied to the team's number, plus equity