Build the ML-powered discovery systems behind what Prime Video shows you

Join Prime Video's Discovery Engineering team to build the low-latency ranking and retrieval systems, offline ML pipelines, and LLM-based evaluation systems that decide what customers see when they open Prime Video. Requires production ML integration experience and comfort owning distributed systems end to end.

Practice this interview

Free · a live voice mock calibrated to this exact role

Start the mock interview

Likely format

Amazon SDE loop: online assessment, phone screen, then onsite/virtual loop of multiple rounds covering coding, system design, and Leadership Principles behavioral interviews.

What this interview tests

Low-latency recommendation serving: ANN retrieval, reranking, ranking systems
Offline ML/data pipelines using LLMs and embeddings
LLM-based quality/relevance evaluation systems
End-to-end ownership from design through production, validated by A/B testing

Common question themes

Design a low-latency personalized recommendation serving system

How would you build an offline pipeline to embed and index content using LLMs

Describe evaluating recommendation quality or relevance with automated methods

Tell me about owning a distributed system component end to end

How do you validate real customer impact of a ranking change (A/B testing)

How candidates describe it

Real Software Development Engineer interview stories — retold from candidates' public write-ups, with sources.

Amazon · SDE I (New Grad, MS CS)Offer

Amazon SDE 1 new grad loop: three online assessments, three interviews, offer in five days

A master's new grad went through Amazon's full three-part online assessment and three 45-minute virtual interviews. Every round opened with Leadership Principles questions before the coding began, and the offer landed within five business days.

Interviewed October 2019 · Seattle, WA, US

Amazon · SDE I (New Grad)Offer

Amazon SDE 1 new grad interview: one OA, four back-to-back onsite rounds, offer

A fresher found the posting through a LinkedIn connection, cleared the OA within a week of applying, and faced a six-hour virtual onsite: four one-hour rounds where every behavioral answer drew three to four follow-up questions.

Interviewed January 2021 · Seattle, WA, US

Amazon · SDE INo offer

Amazon SDE I Interview Experience: Global Selling Team, 2022

A LinkedIn referral bypassed Amazon's usual online assessment for this SDE I candidate, who then worked through a telephonic DSA screen and two technical rounds pairing coding problems with Amazon Leadership Principles questions before receiving no further response.

Interviewed March 2022 · Not specified

View the original posting

All Amazon Software Development Engineer interviews

All Amazon interviews