Senior Inference Engineer — Scale Production AI APIs

AssemblyAI, Inc.; Seattle (WA); On-site; USD 190,000 - 225,000; Full time

Job summary

AssemblyAI is hiring a Software Engineer to turn cutting-edge AI research into products our customers rely on every day. You'll sit on a fast-paced engineering team between our research org and our Applied teams, taking new model capabilities from proof of concept to robust, well-documented APIs serving production traffic at scale.

You’ll design and ship customer-facing APIs, scale inference infrastructure for 1M+ users, and collaborate with researchers, applied engineers, and customers to

Qualifications

• Backend engineering experience building production ML infrastructure and/or customer-facing APIs.
• Comfort operating with ambiguity in timelines and priorities.
• Genuine curiosity about AI/ML to understand model behavior for engineering decisions.
• Strong collaboration with researchers, engineers, and customers to inform product direction.

Responsibilities

• Design and ship customer-facing APIs that expose new model capabilities from prototype through launch and ongoing iteration.
• Scale inference infrastructure to support 1M+ API users.
• Productionize new techniques in collaboration with researchers and apply them to latency, reliability, and ergonomics.
• Work with customers and Applied teams to understand usage and feedback for product and research direction.
• Contribute across the stack: backend services, inference infrastructure, SDKs, internal tooling.

Skills

Backend engineering

ML infrastructure

APIs in production

Collaboration

Job description

AssemblyAI is hiring a Software Engineer to turn cutting-edge AI research into products our customers rely on every day. You'll sit on a fast-paced engineering team between our research org and our Applied teams, taking new model capabilities from proof of concept to robust, well-documented APIs serving production traffic at scale.

You’ll design and ship customer-facing APIs, scale inference infrastructure for 1M+ users, and collaborate with researchers, applied engineers, and customers to