Amazon · Primly Community

Amazon frontend engineer interview: what rounds they run and where javascript depth actually matters

frontend_fran · 4 replies

did the amazon frontend engineer loop for an L5 role on the prime video team in january 2026. i've done loops at two other big tech companies before so i have some comparison points.

amazon's frontend loop is interesting because they don't differentiate as sharply between frontend and fullstack as some companies do. they expect FE engineers to have real javascript depth AND to be comfortable with light system design.

the rounds for my loop:

coding round 1: pure data structures. no DOM, no CSS. lc-style problem on arrays. i was surprised. they treat frontend coding as algorithmic coding, not component-building. if you've been only practicing 'build this react component' you'll be caught off guard.

coding round 2: more applied. a javascript-specific problem about async patterns. they gave me a scenario where i had to implement a rate-limited queue of async requests. this was much more FE-relevant. things like promise chaining, error handling, cancellation semantics. this is where real js depth shows up.

system design: they asked me to design a news feed that needed to handle real-time updates and lazy loading at scale. definitely more fullstack than pure FE. i talked through websocket vs SSE for the real-time layer, virtualized lists for the DOM performance side, and a CDN edge caching strategy. they seemed satisfied with that framing.

LP round: ownership and customer obsession. had solid stories. they pushed on 'how did you know the customer impact' which is a good signal that they want metric-backed answers not vibes.

bar-raiser: this was a mini system design focused on a performance problem. they asked me to diagnose why a page had poor LCP and what i'd change. very concrete. first principles web perf knowledge matters here.

one thing i'd flag: amazon FE interviews are more algorithmic than the industry average for FE. if you come from a company where FE interviews are 80% react and CSS, shift your prep mix. i'd say 60% algo, 30% JS depth, 10% system design for the FE track.

i got the offer. starting q2 2026.

4 replies

corp_refugee

the 'lc-style problem with no DOM' for frontend is genuinely amazon specific. i did a loop there two years ago and had the same experience. good that you flagged it. most FE interview prep resources don't tell you this.

mobile_mara

the rate-limited async queue problem sounds exactly like something i'd get in an ios interview framed differently. makes sense they test that kind of concurrency reasoning for any client-side role.

qa_quinn

playing contrarian here: if amazon FE interviews test mostly algo with a thin JS layer on top, are they actually hiring the best frontend engineers or just the best algo-prepped engineers who happen to know JS? asking because the bar being algorithmic might explain some of the FE quality issues in amazon products historically.

frontend_fran

genuinely fair point and i thought about this. i think the real-world answer is: amazon's frontend at scale has enough infrastructure complexity that algo instincts matter. but i agree a pure frontend artisan who cares about pixel-perfect craft and accessibility depth might not optimize well for this loop.