Data & Analytics Intern

Legora Legora · Vertical AI · Stockholm, Sweden · Corporate

This internship focuses on designing, building, and deploying an internal analytics agent that leverages Legora's data stack. The agent should be able to answer business questions, navigate data models, and improve over time. The role involves end-to-end ownership, from scoping to iteration, with a strong emphasis on shipping a functional product that remains in use post-internship.

What you'd actually do

  1. Design, build, and deploy an internal analytics agent that helps Legora employees explore and understand our data.
  2. Architect the agent so it improves over time — retaining context across conversations, learning from prior interactions, and getting smarter the more it's used.
  3. Integrate the agent with our data warehouse (Snowflake), semantic layer, and dbt models so it can answer questions grounded in real, modelled data.
  4. Build the surrounding infrastructure: evaluation harnesses, observability, guardrails, and a deployment pipeline that lets the agent evolve safely.
  5. Partner with data, product, and engineering to shape what the agent should do — and what it shouldn't.

Skills

Required

  • Currently studying or recently completed a degree in Computer Science, Engineering, Data Science, or a related technical field.
  • You've built and deployed agentic systems before — not just called an LLM API in a script, but designed agents with tools, state, and real production constraints in mind.
  • Comfortable across the stack: you can write production Python, you understand how to model data (dbt, SQL, warehouse design), and you've worked with modern LLM frameworks and orchestration tools.
  • You think like an engineer about reliability, evaluation, and iteration — not just like a prototyper.
  • A bias toward shipping. You'd rather get something crappy in front of users in week two than have something perfect in week ten.

Nice to have

  • Familiarity with at least some of: dbt, Snowflake (or similar cloud warehouses), semantic layers, vector stores, LLM observability tooling, and modern agent frameworks.

What the JD emphasized

  • built and deployed agentic systems before
  • bias toward shipping

Other signals

  • building an internal analytics agent
  • designing and deploying an agentic system
  • own the project end-to-end
  • hands-on, ship-it summer
  • what you build is still in use after you leave