+1 (726) 227-2971
Vistelio

Interviewing Looker Developers in 2026: Interview Questions

Updated August 2026. The 2023 version of this post predates Looker API 3.x removal, BigQuery Editions, Looker (Google Cloud core) and Gemini in Looker. The questions below reflect what a Looker developer needs to know now.

Demand for skilled Looker developers has not gone away; what has changed is the job. A strong Looker developer in 2026 is a semantic-layer engineer: someone who can model well, control warehouse cost, ship LookML through CI, run a core migration and make AI features trustworthy. Here are the questions we use to find that person.

1. Background and experience

  • "Walk me through the most complex LookML project you have owned. How many models and explores, how many developers, and how did you keep it maintainable?"
  • "Which warehouses have you modeled against in Looker? What is different about BigQuery versus Snowflake from the LookML side?"
  • "Have you worked on Looker (original), Looker (Google Cloud core), or both? What changed for you?"

2. Modeling proficiency

  • "Explain the difference between a SQL-based derived table and a native derived table, and when you would persist either one."
  • "A dashboard shows revenue that is double what finance reports. Walk me through how you would find a join fan-out and fix it."
  • "How do symmetric aggregates work, and what do they require from your views?"
  • "When would you use extends, and when would you use refinements?"

3. Performance and cost

  • "We are on BigQuery. How would you decide between on-demand and an Edition, and what would you change in LookML to reduce spend either way?"
  • "Describe how you would introduce aggregate awareness to a dashboard that takes thirty seconds to load."
  • "Explain persist_for versus datagroup_trigger. Why should they not appear together?"
  • "Where in System Activity do you look first when someone says Looker is slow?"

4. Security and access control

  • "Design row-level security for a multi-tenant embed. Which of access_filter, sql_always_where and required_access_grants do you use, and why?"
  • "What is the risk of a user attribute that users can edit themselves?"
  • "How do model sets, permission sets and roles fit together?"

5. API, CI and delivery

  • "Our scripts stopped working in 2023. Why, and how would you migrate them?" (The answer should mention API 3.x removal, API 4.0 and the SDKs.)
  • "Describe a CI pipeline for LookML. What do LookML data tests catch, what does a linter like LAMS catch, and what does SQL validation with Spectacles catch?"
  • "Have you used advanced deploy mode and deploy webhooks? What problem do they solve?"

6. Migration and platform

  • "What changes when an instance moves from Looker (original) to Looker (Google Cloud core)? What breaks first?"
  • "How do you inventory an instance before a migration?"

7. AI and Gemini

  • "We want to roll out Gemini in Looker and Conversational Analytics. What would you change in the model first, and how would you measure whether the answers are right?"
  • "Give an example of a question a natural-language assistant gets wrong on an uncurated model, and the LookML fix for it."
  • "Does Gemini replace the need for a Looker developer?" (We wrote our own answer here.)

8. Collaboration and judgment

  • "How do you handle a stakeholder who wants a metric defined differently from the rest of the company?"
  • "Tell me about a time you pushed transformation logic out of Looker and into dbt, or the reverse. Why?"
  • "How do you keep up with Looker release notes, and what is the most recent change that affected your work?"

Scoring

We score candidates on four things: correctness on the technical questions, the ability to explain trade-offs rather than recite rules, evidence of having operated an instance (cost, CI, security) rather than only building dashboards, and communication. A candidate who answers the modeling and security sections well but has never looked at a BigQuery bill or a deploy pipeline is a strong dashboard developer, not yet a Looker developer for 2026.

An alternative to hiring in house is to engage a specialist firm. Vistelio provides senior, US-based Looker developers as embedded members of your team, for staff augmentation, rescues, migrations and AI rollouts. Get in touch.