Companies hiring inference engineers in 2026
Frontier labs, inference providers, GPU and accelerator companies, big tech and trading firms that post inference, ML systems and GPU kernel roles, with links to each careers page and what each loop publicly documents.
4 MIN READ · UPDATED 16 OCTOBER 2026
PRACTICE THIS:Inference System Design ·Serving Engine Internals
Five kinds of employer, five kinds of loop
Every company in the directory below had a posting for inference, model serving, ML performance or kernel work when we checked in October 2026. They fall into five families, and the family predicts the loop better than the company's size. Frontier labs weight serving system design and implement-from-scratch coding for their own models. Inference providers, whose product is serving other people's models, weight engine internals, scheduling and debugging. GPU and accelerator companies weight kernels, runtimes and profiling on their own silicon. Big tech teams often add compilers and performance analysis. Trading firms put models on the critical path and weight latency and whole-system performance.
Most of these companies publish no inference-specific interview loop, and we say so on each company page rather than guessing. Where a company does publish its process, the page lists the stages with the source.
How to use the list
Search each careers page for more than one title: inference engineer, member of technical staff on an inference team, performance engineer, ML systems engineer, deep learning performance engineer, kernel engineer and model serving engineer all describe this work. Then read the responsibilities for the nouns that matter: serving engines, KV cache, batching, quantization, kernels and latency targets.
The directory: 35 companies, and where to apply
Two links per company: its careers site, since a req there is live by definition, and the page here that documents what its inference loop tests.
Few teams post the title “Inference Engineer”. Serving, kernel and performance work shows up under these names as well:
Frontier model labs
Inference and serving teams for the labs' own models.
Inference providers and AI platforms
Companies whose product is serving models for others.
GPU and accelerator companies
Kernels, runtimes and inference software for their own hardware.
Big tech
Inference and ML performance teams inside large product organizations.
Trading firms
ML performance roles where models sit on the critical path.
No job listings live on this site, and none are checked by us. A req can open, close or change title in a week; the careers link is where to confirm it.
Question topics that test what this guide covers:
FAQ
Frontier labs such as OpenAI, Anthropic, Google DeepMind, Meta, xAI and Mistral; inference providers and platforms such as Fireworks AI, Together AI, Baseten, Modal and Perplexity; GPU and accelerator companies such as NVIDIA, AMD, Cerebras, d-Matrix and Etched; big tech teams at Google, Microsoft, Amazon and Snowflake; and trading firms including Jane Street, Hudson River Trading, Two Sigma and Optiver. The directory lists every company we track, with careers links.