As of October 26, 2026, Level lists 328 live postings that match "inference engineer". Most ask for senior experience, and about one in three allow remote work. OpenAI and Anthropic cut model prices this year, and Level does not show that those cuts create inference jobs. Here is who is hiring today.
What happened
Max Woolf published a timeline of LLM price cuts on October 25, 2026. It says OpenAI cut the price of GPT-5.6 Luna by 80% on July 30.
On September 22, OpenAI announced GPT-6 Sol. It costs $2 per million input tokens, and output tokens cost $10 per million. The post says that halves the price of the discounted GPT-5.6 Sol.
Anthropic made a similar move that day with Claude Opus 5.5. Input tokens cost $4 per million, and output tokens cost $20 per million. The post says it cuts total cost by 40% to 60% against Opus 5.
The post also cites Vercel's AI gateway data. It puts cost per token at 43.3% of the April 2026 level, while total spend keeps rising.
Open models add pressure. The post counts 29 providers hosting the DeepSeek V4.1 Flash 1017 model on OpenRouter, against 13 for the earlier V4 Flash model in May.
The serving side is moving too. On October 22, Cerebras described how it deploys trillion-parameter models on its wafer-scale chips. Its post says it runs multi-trillion parameter models at 1,000 to 2,000+ tokens per second.
Level does not show that price cuts create jobs. The data below only shows who is hiring inference engineers today.
Who is hiring inference engineers?
Level found 328 live postings that match "inference engineer" on October 26, 2026. Of those, 142 have "inference" in the job title.
NVIDIA has the most matches, with 43 postings. Amazon has 36, and Capital One has 24. These are text matches, so some are general infrastructure or machine learning roles.
Anthropic has 14 matching postings. Together AI and Cerebras have 12 each.
Look only at titles that contain "inference" and the order changes. Amazon leads with 16, NVIDIA has 14, and Anthropic has 12.
Together AI and Cerebras then follow with 11 each. See the pages for NVIDIA, Anthropic and Cerebras.
Of the 328 matches, 283 are Builds AI, which is 86%. Works on AI roles make up another 10%. See Level 4 for the full list of roles where building AI is the job.
How senior are inference engineer jobs?
Senior is the norm. Level tags 61% of the postings as senior. Lead roles add another 17%.
Entry points are thin. Internships and junior roles make up only 5% of the matches.
If you are early in your career, read the AI engineer salary guide and compare it with machine learning engineer jobs. Those roles feed into inference work.
What do inference engineer jobs pay?
178 of the postings list pay in US dollars. The median listed minimum is $189,000+.
The median listed maximum is $277,000+. These are posted ranges, not offers, and they exclude equity.
Among the 142 postings with "inference" in the title, 86 list USD pay. The median minimum there is $190,000+, and the median maximum is $287,000+.
What skills do the postings ask for?
Level reads the tools named in each posting. PyTorch appears in 144 of the 328 postings, or 44%.
vLLM appears in 107 postings, or 33%. Kubernetes appears in 54, or 16%, and CUDA appears in 43, or 13%.
The vLLM count rises to 63 among the 142 postings with "inference" in the title. That is 44% of that group.
If you want to learn the serving stack first, start with open source. The vLLM page lists every posting that names it, and our article on vLLM jobs covers the employers.
Can you work as an inference engineer remotely?
Sometimes. 104 of the 328 postings allow remote work, or 32%.
Location is concentrated. 219 postings are in the United States, which is 67% of the matches.
Browse the wider list of remote AI jobs if you can work from anywhere.
Open roles right now
These are the newest live matches. Each links to the full posting.
