Senior Multimodal AI Systems Engineer

Cohere; New York, San Francisco (NY, CA); Hybrid; USD 180,000 - 250,000; Full time

Benefits offered by this job

Weekly lunch stipend

Full health and dental benefits

RRSP matching / 401(k)

Parental leave top‑up

Enrichment benefits stipend

6 weeks paid vacation

Travel budget to other offices

$500 home office stipend

Job summary

Cohere is seeking a technical staff member focused on multimodal AI. You will design and develop systems integrating text, speech, and vision, and conduct research on cutting‑edge compute infrastructure. You will collaborate with world‑class teams to push the boundaries of multimodal representation learning.

The role emphasizes hands‑on engineering, scalable systems, and potential publications; remote‑friendly environment with offices in major cities worldwide.

Qualifications

• Excellent software engineering skills with a track record of scalable systems.
• Proficient in Python and DL frameworks (JAX, PyTorch, TensorFlow).
• Experience with distributed training for large multimodal models.
• Familiar with autoregressive models for multimodal tasks.
• Publications in top venues are a plus.
• Experience writing CUDA kernels is a plus.

Responsibilities

• Design and develop cutting‑edge multimodal AI systems, integrating text, speech, and vision.
• Conduct research and experiments on our compute infrastructure, exploring novel ideas in multimodal representation learning.
• Collaborate closely with world‑class teams, learning from and contributing to their expertise in the field.

Tools

HuggingFace

CUDA kernels

Python

JAX

PyTorch

TensorFlow

Distributed training

Multimodal AI

Publications

Job description

Cohere is seeking a technical staff member focused on multimodal AI. You will design and develop systems integrating text, speech, and vision, and conduct research on cutting‑edge compute infrastructure. You will collaborate with world‑class teams to push the boundaries of multimodal representation learning.

The role emphasizes hands‑on engineering, scalable systems, and potential publications; remote‑friendly environment with offices in major cities worldwide.