Member of Technical Staff, Mle

Cohere Cohere · AI Frontier · San Francisco, CA · Modeling

This role focuses on applying and customizing Cohere's frontier LLMs for enterprise customers, involving post-training, retrieval, and agent integrations. The individual will design and deliver production-ready models, influence the development of foundation models, and operate with significant ownership, combining application, research, and customer-facing engineering.

What you'd actually do

  1. Contribute to the design and delivery of custom LLM solutions for enterprise customers.
  2. Translate ambiguous business problems into well-framed ML problems with clear success criteria and evaluation methodologies.
  3. Build custom models using Cohere’s foundation model stack, CPT recipes, post-training pipelines (including RLVR), and data assets.
  4. Develop SOTA modeling techniques that directly enhance model performance for customer use-cases.
  5. Work as part of Cohere’s customer facing MLE team to identify high-value opportunities where LLMs can unlock transformative impact to our enterprise customers.

Skills

Required

  • Python
  • core ML/LLM frameworks
  • large-scale datasets
  • distributed training or inference pipelines
  • LLM architectures
  • tuning techniques (CPT, post-training)
  • evaluation methodologies
  • meaningfully shape LLM performance

Nice to have

  • broad view of the ML research landscape
  • desire to push the state of the art

What the JD emphasized

  • production-ready models
  • train and customize frontier models
  • customer domains
  • custom LLM solutions
  • post-training
  • retrieval + agent integrations
  • model evaluations
  • SOTA modeling techniques
  • customer use-cases
  • customer-facing MLE team

Other signals

  • customer-facing
  • custom LLM solutions
  • train and customize frontier models
  • influence foundation models