Ruby on Rails Interview Questions to Ask Before You Hire

12 Ruby on Rails interview questions — with what to listen for

1. 1

   Explain the N+1 query problem in Active Record, and your full toolkit for finding and fixing it.

   What a strong answer covers

   Lazy association loading firing per-record queries; includes/preload/eager_load with the differences named, strict_loading as a guardrail, bullet or query-log inspection for detection, and counter caches for counts. Knowing when eager loading is itself the problem (loading the world) rounds out a senior answer.

2. 2

   Where do you put business logic in a mature Rails app — and what is your honest position on fat models, service objects, and concerns?

   What a strong answer covers

   A defensible position held with pragmatism: models for persistence-adjacent logic, POROs or service objects for multi-model workflows, concerns for genuine shared behavior rather than junk drawers, form objects for complex input. Dogma in any direction matters less than consistency and testability.

3. 3

   Walk me through a zero-downtime migration on a large production table — adding a column with a default, or removing one.

   What a strong answer covers

   Multi-step deploys: additive change, backfill in batches, then constraints; ignored_columns before column removal so running code does not crash; awareness of locks, strong_migrations-style guardrails, and index creation concurrently. This is the sharpest single test of production Rails experience.

4. 4

   How do callbacks go wrong in Rails applications, and what is your discipline around them?

   What a strong answer covers

   Hidden side effects firing on every save (emails from after_save, cascading updates), test pain, and ordering surprises. Discipline: callbacks for data integrity close to the record, explicit service-layer orchestration for business side effects, after_commit for anything external. The email-sent-during-backfill war story is the classic.

5. 5

   Design a background-job system for a Rails feature — say invoice generation with a third-party API. What do you enqueue, and how do you make it safe?

   What a strong answer covers

   Active Job over Sidekiq or the newer Solid Queue, idempotent jobs keyed to survive retries, enqueue after commit, small serialized arguments (IDs not objects), retry with backoff and dead-set monitoring, and rate limiting against the API. Idempotency reasoning is the center of gravity.

6. 6

   What is your caching strategy in Rails — which layers do you use and how do you avoid stale-data bugs?

   What a strong answer covers

   Fragment caching with russian-doll key composition, low-level Rails.cache for expensive computations, HTTP caching where applicable, and key-based expiration over manual invalidation. A stale-cache incident story with the key-design fix demonstrates the scars this question is looking for.

7. 7

   How does Hotwire (Turbo and Stimulus) change how you build Rails UIs, and where do you still reach for React?

   What a strong answer covers

   Turbo Drive/Frames/Streams for server-rendered interactivity with minimal JS, Stimulus for behavior sprinkles, and honest boundaries: highly stateful client experiences (editors, dashboards with heavy interaction) may still justify a JS framework island. Current-stack fluency plus proportion, not tribal answers.

8. 8

   A Rails endpoint has p95 of two seconds. Take me through your diagnosis, in order, with the tools you would use.

   What a strong answer covers

   APM or rack-mini-profiler first, then the ladder: N+1s, missing indexes verified via EXPLAIN, over-rendering partials, memory bloat and GC pressure, external calls in-request moved to jobs. Measurement-ordered, with named tools, is the difference from a memorized optimization list.

9. 9

   What does a safe Rails-version upgrade look like for an app two major versions behind?

   What a strong answer covers

   One version at a time, dual-boot testing (next_rails style), deprecation-warning burndown first, load_defaults stepped carefully, gem compatibility audit, and a strong test suite as the precondition — building one where absent. Rewrite pitches or 'jump straight to latest' fail the question.

10. 10

    How do you test a Rails application — your stack, what gets which kind of test, and how you keep the suite fast?

    What a strong answer covers

    RSpec or Minitest with conviction, model and request specs as the backbone, system specs sparingly for critical flows, factories with restraint (build over create, traits), and speed discipline: parallel testing, profiling slow specs, avoiding needless persistence. Factory-cascade bloat named as a smell is a strong signal.

11. 11

    How do you keep a Rails app secure beyond the framework defaults — what do you audit for?

    What a strong answer covers

    Strong parameters and mass-assignment review, authorization coverage (Pundit/CanCanCan) with a test for every policy, unscoped-find IDOR checks, brakeman in CI, dependency audit, and secrets via credentials or vault. Naming IDOR as the most common real-world Rails hole shows audit experience.

12. 12

    You inherit a 300k-line Rails monolith with 40-minute CI and teams blocking each other. What are your options, and what do you actually do first?

    What a strong answer covers

    Resist the microservices reflex: modularization first (packwerk-style boundaries, engines), CI parallelization and test-suite surgery, ownership maps, and extraction only for genuinely independent domains with the operational cost priced. Sequencing and organizational awareness are what make this answer senior.

Frequently asked questions