The Framework Trap
Most frontend interviews over-index on framework knowledge — React hooks, Vue reactivity, Angular change detection. Framework fluency is a 2-week learning curve for a strong engineer; evaluating it as a primary signal means you're filtering for people who recently used your specific stack, not people who build great products.
Evaluate the underlying skills: how they think about component design, state management tradeoffs, performance at scale, and user experience quality.
Core Evaluation Dimensions
UI Craft: Does their work look and feel considered — not just functional? Review prior work or portfolio. This is one of the rare roles where portfolio evidence is more reliable than interview performance.
Performance Judgment: Do they understand what actually makes interfaces feel fast? Candidates should be able to discuss the critical rendering path, lazy loading strategy, and when not to optimize without prompting.
Accessibility Awareness: Treating a11y as an afterthought is a strong signal of low product care. Ask what accessibility means to them in practice, not in theory.
Cross-Functional Collaboration: Strong frontend engineers have an opinionated relationship with design. Look for evidence of productive design-engineering dialogue, not just pixel-perfect implementation.
Portfolio Review as a Primary Signal
Ask candidates to walk through something they built and are proud of. Strong frontend engineers have opinions about the work — what they'd change, what constraints they navigated, where they pushed back on design.
Candidates who can't discuss their work critically either didn't own it or don't reflect on it. Both are risks for a role that requires continuous iteration.
What to Avoid
JavaScript trivia questions test cram-ability, not engineering quality. 'What is the difference between null and undefined' tells you nothing about whether someone builds good products.
Live-coding under observation disadvantages engineers who work reflectively rather than performatively. Consider code review or take-home formats that better simulate actual work conditions.