What senior frontend interviews actually test in 2026

The first draft of Gapmap's frontend syllabus came from personal experience, a folder of prep notes. We threw it away. Personal experience is one interview loop wide, and a syllabus needs more than that, so we rebuilt it from three kinds of sources and kept only what they agree on:

1. roadmap.sh: the industry-canonical skill trees (frontend, JavaScript, React, performance), with their own must-know flags.
2. The interview handbooks: Frontend Interview Handbook, GreatFrontEnd's interview and system-design playbooks, Tech Interview Handbook: round types and per-round topic taxonomies.
3. Real loop evidence from 2025–26: company loop guides, first-person senior interview reports, aggregate question collections.

These sources have different authors and different incentives, so where they agree the signal is trustworthy. Here is where they agree.

The shape of a frontend loop

A senior frontend loop in 2026 looks the same across sources: a knowledge quiz, a JavaScript coding round, a UI coding round, a frontend system design round, and behavioral. Two things about this list matter more than the rest.

System design is the seniority gate. Every source family says the same thing: a weak system-design round approximately equals rejection at senior level. It is the round that decides your level.

LeetCode-style DSA is declining for frontend roles. Most companies have replaced it with practical JS and UI coding. The notable exception is Meta-style loops, where algorithmic rounds persist. If you're not targeting those, grinding graph problems is misallocated prep time.

The cluster everyone asks

One cluster of topics appears in nearly every source. If your prep time is limited, this is the priority list:

• JavaScript internals: the event loop, closures, this and prototypes, promises and async control flow, type coercion. Equality algorithms get flagged as "advanced" but come up more than people expect.
• React's rendering model and hooks: why a component re-renders and what hooks guarantee, rather than API trivia.
• State management with a justified choice. Nobody asks "do you know Redux" anymore. The question is "classify this state, pick a tool, defend the pick": server state vs client state, Context vs a store, optimistic updates.
• Performance diagnosis. "The app takes 8 seconds to load, what do you do?" is near-universal, and the expected answer starts with measurement (Lighthouse, Core Web Vitals), not with a memorized list of tricks.
• Frontend system design: architecture layers, data flow, pagination, caching and invalidation, race conditions.

What separates senior answers

The sources also converge on how seniors are distinguished from middles, and it's rarely more trivia:

• Production reality. Loading states, error states, race conditions, empty states. Handling the stuff that doesn't appear in tutorials.
• Accessibility and internationalization, named as senior evaluation axes in real loop reports, and rising.
• Driving scope: asking clarifying questions, stating trade-offs, cutting the right corners out loud.

One finding surprised us. Real loops increasingly include code-review rounds ("here's a PR, review it") and bug-bash formats, and almost no prep material covers them. What gets practiced and what gets tested have drifted apart.

What we left out

The evidence also says what not to prep. We excluded tool-brand trivia (webpack-vs-vite internals; concepts outrank brands in every source), backend-style capacity estimation (explicitly out of scope in frontend system design), and DSA grinding (see above). Testing showed up as "occasional"; it's on our watchlist, but it didn't make the core 34 topics.

What to do with this

Our product is built on this analysis: the topic map you get after the diagnostic is this syllabus with your actual level measured against it. Even if you never sign up, the priority order above is yours to use. Internals cluster first, system design as the senior gate, performance as diagnosis rather than tricks.

The diagnostic takes about 15 minutes and is free. It will tell you which of these topics are already fine, and which ones would have failed you last week.