Inference Engineer

techire.®; San Francisco (CA); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Medical insurance (including dental &视

Dental insurance

Vision insurance

401(k)

Relocation support

Immigration support

Meals in office

Job summary

techire.® seeks engineers to build the infrastructure that runs next-generation multimodal foundation models at scale. You will design and implement real-time inference pipelines and distributed systems to support low latency, high reliability production workloads.

You’ll collaborate with researchers to productionise new model architectures and drive ownership from day one, focusing on end-to-end reliability and observability across the stack.

Qualifications

• Strong software engineering fundamentals for large-scale systems.
• Experience with ML inference pipelines or serving generative models in production.
• Ability to work through ambiguous technical challenges and deliver zero-to-one systems.
• Experience implementing modern machine learning research into production environments.

Responsibilities

• Build low-latency inference and serving infrastructure for foundation models across Transformers, SSMs and hybrid architectures.
• Design scalable, reliable distributed systems that support production AI workloads.
• Develop monitoring and observability across the inference stack.
• Work closely with researchers to productionise new model architectures.
• Help shape technical direction with significant ownership from day one.

Skills

Distributed systems

ML inference

Production engineering

Ambiguity tolerance

Research-to-prod translation

Tools

CUDA

Triton

vLLM

SGLang

Continuous Batching

Job description

Help build the inference stack behind the next generation of multimodal foundation models.

Most inference roles are about making existing models run faster.

This one is about helping define how entirely new model architectures are served at scale.

You’ll be building real-time multimodal AI capable of processing enormous streams of text, audio and video. The research is pushing beyond today’s transformer limitations, and your work will make those models usable in production.

You’ll sit between frontier research and product engineering, designing the infrastructure that allows cutting-edge models to run with low latency, high reliability and at scale. If you enjoy solving systems problems where every millisecond matters, you’ll feel at home here.

Your focus

• Build low-latency inference and serving infrastructure for foundation models across Transformers, SSMs and hybrid architectures.
• Design scalable, reliable distributed systems that support production AI workloads.
• Develop monitoring and observability across the inference stack.
• Work closely with researchers to productionise new model architectures.
• Help shape technical direction with significant ownership from day one.

You’ll bring

• Strong software engineering fundamentals and experience building large-scale distributed systems.
• Experience with ML inference pipelines or serving generative models in production.
• The ability to work through ambiguous technical challenges and deliver zero-to-one systems.
• Experience implementing modern machine learning research into production environments.

Experience with vLLM, SGLang, Continuous Batching, CUDA or Triton would be particularly valuable but isn’t essential.

Alongside compensation, you’ll receive fully covered medical, dental and vision insurance, 401(k), relocation and immigration support, plus daily meals in the office.

If you’re interested in building the infrastructure that enables the next generation of AI models to run in real time, we’d love to tell you more.