Data Scientist, North Insights

Cohere Cohere · AI Frontier · United States · Agentic Platform

This role focuses on building and shipping customer-facing analytics and predictive models for an enterprise AI company. The Data Scientist will drive product strategy, measure AI impact, and build LLMs for data enrichment, ultimately delivering insights and value to enterprise customers. The role involves end-to-end ownership from framing questions to shipping production models.

What you'd actually do

  1. Build bleeding-edge agentic analytics: Agentic analytics is far from a solved problem, and we want to be the company that solves it. We need the sharpest minds with the curiosity, drive, and focus required to build the tools necessary to bring order and clarity to real world data.
  2. Define AI impact measurement: own the end-to-end analytics story around adoption, growth, and impact that North leadership uses to prioritize features by measured value rather than customer feedback alone.
  3. Ship customer-facing analytics: insights that our enterprise customers recognize as evidence of value created.
  4. Build small language models to categorize, classify, and enrich message data to give customers visibility into how their teams are using our products in their enterprise.
  5. Design and run experiments: A/B tests, causal inference studies, and opportunity sizing that directly map to product and go-to-market decisions.
  6. Build predictive models that matter: develop and deploy models for forecasting, segmentation, propensity scoring, and opportunity sizing across Cohere's core business lines.
  7. Act like an owner: no waiting around. You'll define analytical priorities within your scope, push initiatives from question to production, and own what you build.

Skills

Required

  • Strong command of SQL, Python, and Git
  • expertise in statistical inference, experimental design, and predictive modeling
  • A proven ability to turn ambiguous business questions into rigorous analytical problems, with clear and compelling recommendations to match
  • Experience building and deploying production models, not just notebook analysis. You know what it takes to get a classifier down the stack from prototype to production volume.

Nice to have

  • modern data stack tools such as BigQuery, dbt, Looker, or Airflow
  • Comfort operating in ambiguity, managing multiple workstreams across different stakeholders, and distilling insights into a concise, actionable narrative
  • A collaborative style grounded in empathy as much as in data

What the JD emphasized

  • build bleeding-edge agentic analytics
  • agentic analytics is far from a solved problem
  • build small language models
  • build predictive models

Other signals

  • building analytics that shape product and company strategy
  • designing experiments that prove or kill our biggest bets
  • helping enterprise customers understand what foundational AI actually means for their bottom line
  • shipping insights, tools, and results that product leaders, sales teams, and enterprise customers rely on
  • build small language models to categorize, classify, and enrich message data
  • build predictive models that matter