Adobe ML/Research Intern Interviews: 12 Experiences (Part 4 of 4)

Shared by janebeingjane ·
Company
Adobe
Level
Internship
Stages
Phone screen

Seven Adobe ML and research intern reports (Dec 2023 to Oct 2024) mix easy-to-medium coding, ML basics and project questions; one asked an ETA design.

At a glance

  • Process: ML interns had an online test of linear algebra programming where reported (ML 01; in Python with NumPy for ML 03), then one or two interviews mixing coding, ML theory and projects; the research intern had one 45-minute call with no editor (RI 01).
  • Asked in these reports: coding in four of six ML reports (ML 02, ML 03, ML 05, ML 06), ML fundamentals in four (ML 03, ML 04, ML 05, ML 06), and applied ML or design in four (ML 01, ML 02, ML 05, ML 06).
  • Difficulty: the coding was easy to medium in every report that rated it (ML 02, ML 03, ML 05); ML 03's was a medium binary-search variation, with hints. RI 01 called its online test easy to medium.
  • Practise first: Estimate Trip Time From a Table of Past Trips (ML 01) and Reverse a Singly Linked List (RI 01).

Coverage: 12 interns outside general software engineering: eight machine learning, two research, one web engineering and one technical consulting, in the United States and India where a place is given. Seven are current (Dec 2023 to Oct 2024); five from Jan 2022 to Oct 2023 sit in History.

In this series

Roles in this part: machine learning interns (ML 01–06, ML H01–H02) · research interns (RI 01, RI H01) · web engineering intern (WEB H01) · technical consultant intern (TC H01).

Machine learning interns

What came up

Across the six current ML intern reports, four had a coding part and four asked ML fundamentals such as class imbalance or overfitting. One named a specific ML design case (ML 01, arrival-time estimation); two others had an unnamed pipeline design or case study (ML 05, ML 06), and ML 02 re-solved its coding problem with ML.

6 detailed candidate timelines, loops, and screens

ML 01: ETA modelling and why this internship

Interview date: Oct 2024 · Role: Machine Learning Intern · Location: San Jose · Result: No offer

  1. The online test had no DSA, only linear algebra programming.
  2. One final round: why this internship and what you hope to learn, teamwork, then an ML design for estimating arrival time from the features of a given dataset (Estimate Trip Time From a Table of Past Trips).

ML 02: the same problem, solved twice

Interview date: Oct 2024 · Role: Machine Learning Engineer Intern · Location: San Jose · Result: No offer

  1. About an hour with one interviewer: an easy-to-medium array problem worked with pointers (not named), then the same problem solved with machine learning: approach, data handling and algorithms.

ML 03: a 75-minute panel

Interview date: Oct 2024 (about three weeks after an early-October online test) · Role: Machine Learning Engineer Intern, Summer 2025 · Location: United States · Result: Offer

  1. Online test: three Python questions on linear algebra with NumPy; two fully solved, the third missing one test.
  2. A panel of three planned 30 minutes of coding, 30 of ML and 15 of behavioral. The coding was a medium variation of a well-known binary-search problem, with hints. The ML part probed the candidate's ranking project: LambdaMART over RankNet, NDCG, residual connections and in-context learning. Behavioral was skipped for time.

ML 04: imbalanced classes

Interview date: Jan 2024 · Role: Machine Learning Engineer Intern · Location: United States · Result: No offer

  1. A walk through past work with follow-ups, then what to do when the classes in a classification problem are imbalanced.

ML 05: overfitting, then a pipeline

Interview date: Dec 2023 · Role: Machine Learning Intern · Location: San Jose · Result: Offer accepted

  1. First interview: one easy coding question, then basics of regression, classifiers and transformers, including overfitting versus underfitting and how to fix each.
  2. Second interview: design a full ML pipeline for a use case.

ML 06: binary classification basics

Interview date: Dec 2023 · Role: Machine Learning Engineer Intern · Location: San Jose · Result: No offer

  1. First interview: coding plus binary classification questions. The second covered ML projects and a case study.

History (before Nov 2023)

ML H01: rapid-fire metrics

Interview date: Oct 2023 · Role: Machine Learning Engineer Intern · Location: United States · Result: No offer

  1. Research follow-ups, then quick questions: how an ML pipeline works, the layers in a CNN, and the right metric when wrong positives and wrong negatives both cost something.

ML H02: three 30-minute sessions

Interview date: Apr 2023 · Role: Machine Learning Intern, computer vision · Location: San Francisco · Result: No offer

  1. A hiring-manager call on projects, then a 1.5-hour block with three ML engineers: a simple array question and resume projects first, then two sessions of deep project questions.

Research interns

What came up

RI 01, the only current research intern report, covered two probability questions, list reversal in pseudocode and one project from the resume.

1 detailed candidate timeline

RI 01: reverse a list in pseudocode

Online test: 17 Jul 2024 (interview date not given) · Role: Research Intern, Adobe Research (on campus) · Location: India · Result: Offer

  1. Online test: two basic DSA questions and 16 multiple-choice questions on linear algebra, probability, statistics and aptitude, easy to medium by the candidate's account.
  2. One 45-minute video interview with no editor: two quick probability questions; reverse a linked list, then a doubly linked list, as logic and pseudocode (Reverse a Singly Linked List); and close questions on one project from the resume.

History (before Nov 2023)

RI H01: long-tail data

Interview date: Jan 2022 · Role: Research intern, computer vision · Location: United States · Result: No offer

  1. No coding. One question named: how to handle a long-tail distribution.

Web engineering intern

History (before Nov 2023)

WEB H01: JavaScript, DOM events and React

Interview date: Dec 2022 · Role: Web Engineer Intern (master's student) · Result: Offer

  1. A 45-minute online test with one medium problem, then an engineering manager call on the resume.
  2. Back-to-back rounds: an easy-to-medium problem in JavaScript, then a second problem built on event listeners and DOM changes, and JavaScript basics; a React deep dive; testing tools and behavioral questions.
  3. A same-day recruiter call, mostly behavioral, ended with a verbal offer.

Technical consultant intern

History (before Nov 2023)

TC H01: a JavaScript and React domain round

Interview date: Jan 2022 · Role: Technical Consultant Intern, SheCodes (on campus) · Result: Selected

  1. Online test: domain multiple-choice questions, one coding problem (shorten a string by removing duplicates), an essay and a game-style test.
  2. A 50-minute domain interview (the candidate chose React web): projects, var/let/const scope with output questions, == versus ===, hoisting, the temporal dead zone, currying, higher-order and arrow functions and this, setTimeout with 0, promises and fetching data; React routing, hooks, state versus props, the virtual DOM, function versus class components; HTTP versus HTTPS.

Keep practicing

In this series

Sources: Anywhere useful interview info can be found.

0 total reactions. Your reactions: none.

Comments

0

No comments yet — be the first to share what you know.