Skip to content

Explore

Questions Companies Concepts Guides Sign in

How it works Questions Companies Concepts Guides Pricing Sample scorecard

Interview questions

Find a company or topic. Read the sources, then practice.

Coding study plans Companies Design mock

Loading the question library…

Practice system design interviews out loud with an AI interviewer, a shared whiteboard and a scorecard for your level.

How it works Questions Concepts Questions by company Comparisons Technologies Guides Cheat sheet Sample scorecard FAQ Use it as an app Sign in Pricing About and how we write Terms and guidelines Privacy Policy Contact

© 2026 SysRound. SysRound is an AI practice tool. The AI interviewer and the AI feedback can be wrong. Use it for your own practice, not to hire people.

Search questions

CompanyType a name, then choose a company from the suggestions.

Question type

Difficulty

TopicType a topic, then choose one from the suggestions.

696 questions

Showing 1–25

Machine learningHard
MNIST Loss Bounds per Example and an Adaptive Calibration Error (ACE) Implementation
Companies: OpenAI
ML Coding · Model Evaluation
Last reported Nov 2026
Open workspace
Machine learningHard
Point light source on a line: sample hits, check the density, recover the source by MLE
Companies: OpenAI
ML Coding · Sampling & Estimation
Last reported Oct 2026
Open workspace
Machine learningHard
Majority-Vote Labels from Three Noisy Annotators, Then Beat the Baseline
Companies: OpenAI
Annotation & Label Quality · ML Coding

Page 1 of 28

Last reported Oct 2026

Machine learningHard

Debug a Small PyTorch Transformer with Four Bugs, Then Add a KV Cache

Companies: OpenAI

ML Coding · Transformers & Attention

Last reported Sep 2026

Machine learningHard

Hand-Written Gradients: Prefix-Product Backward Pass and Stable Cross-Entropy

Companies: OpenAI

Arrays & Strings · ML Coding

Last reported Sep 2026

Machine learningHard

Vectorized 1-Nearest-Neighbor in NumPy and Its Equivalence to a Linear Layer

Companies: OpenAI

ML Coding

Last reported Sep 2026

Machine learningHard

Implement a minimal reverse-mode autograd engine on PyTorch tensors

Companies: OpenAI

ML Coding · Neural Networks & Training

Last reported Sep 2026

Machine learningHard

Implement a 1-Nearest-Neighbor Classifier with a Manhattan-Distance Follow-Up

Companies: OpenAI

ML Coding

Last reported Sep 2026

Machine learningMedium

Debug a Transformer Implementation and Implement a KV Cache for Decoding

Companies: OpenAI

ML Coding · Transformers & Attention

Last reported Sep 2026

Machine learningMedium

Identify the Most Important AI Safety Risk

Companies: Anthropic

Last reported Aug 2026

Machine learningMedium

Detecting and Mitigating Harmful Bias in an AI System: A PM's Plan

Companies: OpenAI

AI Product Management · Fairness & Bias

Last reported Aug 2026

Machine learningMedium

Vectorize One-Nearest-Neighbor and Express It as a Neural Forward Pass

Companies: OpenAI

ML Coding · Neural Networks & Training

Last reported Aug 2026

Machine learningMedium

Implement Masked Cross-Entropy with Label Smoothing

Companies: OpenAI

ML Coding

Last reported Jun 2026

Machine learningMedium

Implement 1NN with NumPy

Companies: OpenAI

ML Coding

Last reported May 2026

Machine learningMedium

Compute entropy and implement 1-NN

Companies: OpenAI

ML Coding

Last reported Apr 2026

Machine learningMedium

Forecasting the Generative AI Skills Product Teams Will Need Most in the Coming Year

Companies: OpenAI

AI Product Management

Last reported Apr 2026

Machine learningMedium

Derive Sharded Matrix Multiplication and Backpropagation

Companies: OpenAI

ML Coding · Neural Networks & Training

Last reported Apr 2026

Machine learningMedium

Design a Double Descent Experiment

Companies: Anthropic

ML Fundamentals · Regression & Statistical Modeling

Last reported Apr 2026

Machine learningMedium

Defend a Research Direction and Experiment Design

Companies: OpenAI

Model Evaluation · Project Deep Dive

Last reported Apr 2026

Machine learningMedium

Hallucination Playbook for a Live LLM Product: Measurement, Safeguards and UX

Companies: Anthropic

AI Product Management · LLM Applications

Last reported Apr 2026

Machine learningMedium

How Context Window Length Shapes LLM Quality, Cost and Latency

Companies: OpenAI

AI Product Management · Transformers & Attention

Last reported Apr 2026

Machine learningMedium

Choosing Between Prompting, Retrieval and Fine-Tuning to Specialize an LLM

Companies: Anthropic

AI Product Management · Fine-Tuning & Alignment Techniques

Last reported Apr 2026

Machine learningMedium

Pinning Down What an LLM Hallucination Is: Types, Causes and Boundaries

Companies: OpenAI

AI Product Management · LLM Applications

Last reported Apr 2026

Machine learningMedium

Keeping LLM Output Safe: Dual-Use Requests, Jailbreaks and Layered Defenses

Companies: Anthropic

AI Product Management · AI Safety & Alignment

Last reported Apr 2026

Machine learningMedium

Risks of Anthropomorphizing Large Language Models in AI Products

Companies: Anthropic

AI Product Management

Last reported Apr 2026

Interview questions by company and topic | SysRound