Staff Software Engineer, Agentic Analytics

Cohere•San Francisco, CA
•$250,000 - $325,000•Remote

About The Position

Cohere is hiring a hands-on founding lead for Agentic Analytics, a new team in its Analytics and Data Insights organization. The role owns the end-to-end development of an internal agentic analytics system that answers business questions over governed data, provides the source, query, and freshness behind every answer, and asks for clarification or declines when evidence is insufficient. The system is in pilot on a BigQuery warehouse, supported by tested ingestion pipelines and dbt models. The first-year objective is to bring it into production across Cohere one domain at a time and make onboarding new domains repeatable through configuration, evaluation sets, and minimal new code. The Staff Software Engineer will set the architecture, write production code and evaluations, and lead a small founding team as it grows. The role combines building reliable and secure agent services with defining evaluation standards, engineering the context layer, and coordinating domain rollouts. It works closely with data platform, model evaluation, security, and business domain teams, and reports to the Senior Director of Data Analytics.

Requirements

  • 10+ years building and operating production analytics or software systems, including technical leadership of a team or systems spanning teams.
  • Experience shipping LLM agents to production, including tool use, MCP or similar, retrieval and context engineering, guardrails, and observability.
  • Hands-on experience building natural-language or agentic analytics systems, such as text-to-SQL, analytics agents, or question answering over a semantic layer, that real users relied on, with evidence of correctness.
  • Depth in evaluating LLM systems, including golden sets, result-level comparison, abstention metrics, error taxonomies, and regression testing in CI.
  • Strong Python and SQL skills.
  • Experience designing and operating production services and APIs on a major cloud platform.
  • Fluency with cloud data warehouses such as BigQuery or Snowflake, dbt, and semantic layers such as MetricFlow, LookML, or Cube; ability to reason about data grain, metric definitions, and approved joins.
  • Experience building identity-aware access for agents so they act with the requesting user's permissions and queries leave an audit trail.
  • Experience hiring and developing engineers, with an empathetic leadership style.
  • Genuine interest in AI research and willingness to work hands-on with the technology.
  • Comfort operating amid ambiguity and leading through partners who are not direct reports.

Nice To Haves

  • Experience with Google Cloud.
  • Experience with BigQuery, Snowflake, or similar cloud warehouses.
  • Experience with MetricFlow, LookML, Cube, or similar semantic layers.
  • Experience with MCP or equivalent agent-facing data tools.

Responsibilities

  • Own the end-to-end architecture that turns business questions into governed queries and verifiable answers, including question interpretation, context retrieval, approved definitions and joins, user permissions, and evidence attached to answers.
  • Design core services and write much of the initial production implementation, including agent-facing data tools (MCP or equivalent), context retrieval, analytics agent skills, and answer-serving APIs.
  • Ensure production services meet standards for reliability, latency, cost, security, and observability, and integrate them into places where users work, such as Slack, North, IDEs, and other agents.
  • Build and operate the evaluation harness, define correctness standards for each domain, create ground-truth and held-out question sets (including clarification and decline cases), define result-level scoring and error measures, and run regression tests for changes to context, skills, or models.
  • Decide whether domains are ready for production and prevent releases that do not meet the evaluation bar.
  • Develop the agent context layer from metric definitions, approved joins, code-generated table documentation, business glossary content, and worked examples; turn corrections into reviewed definitions and regression tests.
  • Roll out the system across domains such as marketing, revenue, finance, and product, establishing owners, certified definitions, and domain-specific evaluation sets; measure adoption, accuracy, turnaround, and correction effort.
  • Package tools, skills, context templates, and evaluations, and document them so other Cohere teams can operate the system independently.
  • Hire and lead the team's first engineers, set the technical bar, review work, and develop team members.
  • Partner with data platform, model evaluation, security, and domain owners on warehouse and semantic models, evaluation methodology, access controls, definitions, and rollout.
  • Write design documents and evaluation reports to support decisions across teams and lead effectively through ambiguity and cross-functional partners.

Benefits

  • Weekly lunch stipend of $75/£75 or equivalent in local currency.
  • Full health and dental benefits, including a separate mental health budget.
  • RRSP matching, 401(k), or pension scheme, depending on location.
  • 100% parental leave top-up for up to six months for either parent.
  • Annual enrichment benefits for arts and culture, fitness and wellness, quality time, and workspace improvements.
  • Education and learning stipend for conferences, courses, and coaching.
  • Six weeks of paid vacation (30 working days).
  • Travel budget for remote employees to visit other offices and an annual company offsite.
  • For office-based employees: daily lunch program, snacks, and community and social events.
  • Co-working benefit for employees who are not near an office.
  • $500 home office stipend.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service