practice a little every day, paws crossed ♡
ML System Design Interview Questions
Practice 9403 coding, system design and machine learning questions from 546 companies, sorted by role, topic and interview round. Every question has step-by-step hints, a code runner that supports 8 languages and an AI mentor; members also get the full solution walkthrough.
- 9403
- Questions
- 546
- Companies
- 11
- Topics
- 8
- Languages
Showing 31–60 of 186 results
- AnthropicHardSoftware EngineerEstimate VRAM and compare model parallelism
This question tests the ability to calculate memory overhead for transformer model parameters and activation buffers, while evaluating the architectural trade-offs between tensor and pipeline parallelism for large-scale training.
ML System DesignOnsiteDistributed SystemsGPU ArchitectureTransformer Training - AnthropicMediumSoftware EngineerPerformance-Model a Sharded Matrix Multiply: FLOPs, Memory, Network and Roofline
Model the performance of a matrix multiplication by hand: count floating-point operations, memory bytes and network bytes, apply the roofline model to decide what bounds it, and compare sharding along each dimension across several accelerators. It tests arithmetic intensity, collective communication costs and quantitative reasoning.
ML System DesignOnsite - AnthropicHardSoftware EngineerReview an inference API design for scale
This question tests a candidate's ability to critically evaluate ML system designs for production-scale inference APIs, covering multi-tenancy, GPU resource constraints, and streaming token delivery. It assesses architectural reasoning across reliability, latency trade-offs, and capacity planning in the ML Systems domain.
ML System DesignOnsite - AppleMediumMachine Learning EngineerDesign App Store search
This question evaluates a candidate's ability to design scalable, low-latency app marketplace search systems, covering competencies in query understanding, metadata ingestion and indexing, candidate retrieval and ranking, personalization, multilingual and typo-tolerant handling, cold-start and spam prevention, as well as online experimentation and monitoring. Commonly asked in the ML system design/search engineering domain to probe architectural trade-offs, measurement of product requirements and success metrics, and operationalization, it spans both high-level conceptual reasoning and practical application details.
ML System DesignOnsite - AppleMediumMachine Learning EngineerDesign a grounded voice assistant
This problem evaluates your ability to design a retrieval-augmented generation (RAG) system for a voice assistant, focusing on grounding model outputs in trusted sources to prevent hallucinations.
ML System DesignOnsiteRAGLLMSystem Design - AppleMediumMachine Learning EngineerOptimize image filters on device
This question evaluates competency in on-device machine learning and performance engineering, covering algorithm choice, numerical precision trade-offs, hardware utilization (CPU/GPU/accelerator), memory layout, tiling/streaming, and trade-offs among latency, memory overhead, and battery impact.
ML System DesignOnsite - AtlassianMediumMachine Learning EngineerDesign Jira bug-to-team classification system
This question tests the ability to design an end-to-end ML pipeline for classifying incoming Jira tickets to appropriate engineering teams based on textual and historical metadata.
ML System DesignOnsiteNLPClassificationScalability - AtlassianMediumMachine Learning EngineerDesign a Jira+Confluence RAG assistant
This question evaluates competencies in designing retrieval-augmented generation systems, including ingestion and indexing, retrieval and ranking strategies, prompt-based generation and citation handling, permission-aware access control, and monitoring for freshness and latency.
ML System DesignOnsite - AtlassianHardMachine Learning EngineerDesign an ML-powered search system
This question tests the ability to design a scalable, low-latency, and relevant search system for internal enterprise documents, focusing on the integration of hybrid search (keyword + semantic) and ranking strategies.
ML System DesignOnsiteInformation RetrievalRankingEmbeddings - AtlassianMediumMachine Learning EngineerDesign an internal data-access chatbot
This question evaluates a candidate's ability to design production-ready ML systems that integrate retrieval-augmented generation, secure access control, tool-based workflows, data privacy, auditability, and low-latency interaction.
ML System DesignOnsite - ByteDanceMediumSoftware EngineerDesign a Content Moderation Platform
This problem evaluates your ability to design a scalable, multi-modal content moderation system that filters user-uploaded media using both automated ML inferences and human review workflows.
ML System DesignOnsiteSystem DesignMLOpsScalability - ByteDanceMediumSoftware EngineerDesign a Short-Video Recommendation System
This problem tests the ability to design a multi-stage recommendation pipeline that balances real-time personalization, system latency, and content diversity.
ML System DesignOnsiteSystem DesignRecommendation SystemsScalability - C3 AIHardData ScientistReduce Fresh-Food Waste with an End-to-End ML Approach
This problem tests the ability to design an integrated ML system that moves beyond simple demand forecasting by addressing inventory dynamics, supply constraints, and decision optimization.
ML System DesignOnsiteSupply ChainOptimizationCensored Data - Capital OneMediumMachine Learning EngineerDesign model deployment, monitoring, and low-latency inference
This question evaluates competency in ML system design and production engineering—covering model deployment and versioning, safe rollouts and rollbacks, monitoring of service health, data quality/drift, model performance and business metrics, and latency optimization for low-latency inference.
ML System DesignOnsite - CohereMediumMachine Learning EngineerDesign How an LLM Detects Questions Beyond Its Training Cutoff and Says So
Design how a large language model recognizes questions that depend on information after its training cutoff and tells the user where its knowledge ends. The question tests data construction for calibrated abstention, prompt versus fine-tuning versus retrieval trade-offs, and evaluating both over-refusal and hallucination.
ML System DesignOnsite - CohereMediumMachine Learning EngineerDesign a Post-Training Pipeline to Improve a 7B Model's Coding for an Enterprise
Design an end-to-end post-training pipeline that improves a 7B language model's coding ability for an enterprise customer, covering data collection, SFT, RLHF, evaluation, serving and a production feedback loop. The question probes dataset construction, reward model choice and why RLHF is still needed when SFT can learn from preference data.
ML System DesignOnsite - CohereMediumMachine Learning EngineerImprove LLM reasoning for a domain task
This problem tests your ability to design a robust pipeline for an LLM-based system that requires multi-step reasoning in a specialized domain (e.g., technical legal analysis or medical diagnosis).
ML System DesignOnsiteLLMReasoningRAG - CreditkarmaMediumMachine Learning EngineerDesign a Revenue Ranking Platform
This problem tests your ability to design a multi-stage ranking system that balances business revenue goals with user relevance, specifically addressing sparse conversion data and long-term user retention.
ML System DesignOnsiteRankingRecommendationMulti-task Learning - CrestaMediumSoftware EngineerDesign an Evaluation Platform for AI Agents Across Prompt and Model Changes
An ML system design question about a platform that lets teams evaluate AI agents and measure how a change of system prompt or model affects quality, tool use and response speed. It tests evaluation design, scoring with references and LLM judges, statistical comparison of versions and scalable execution.
ML System DesignOnsite - DatabricksMediumMachine Learning EngineerDesign Harmful Content Detection
This problem tests the ability to design an end-to-end machine learning system for multimodal harmful content detection at scale, focusing on pipeline architecture, latency, and data governance.
ML System DesignOnsiteSystem DesignMultimodalScalability - DatabricksMediumMachine Learning EngineerDesign Harmful Content and OOM Detection
This question tests the ability to design an end-to-end ML system for real-time safety, covering taxonomy, data strategies, model architecture, and operational monitoring.
ML System DesignOnsiteSystem DesignMLOpsSafety - DatabricksMediumSoftware EngineerDesign Retrieval for a Coding Agent: What to Index, Handling Noise, and Metrics
Design the retrieval layer that feeds code and documentation to an LLM coding agent, deciding what to index, how to split and rank it, how to handle noise, and which metrics prove it works. The question tests code-aware indexing, hybrid lexical and dense retrieval, reranking, freshness, and offline and end-to-end evaluation design.
ML System DesignOnsite - DatabricksMediumMachine Learning EngineerDetect Harmful Content in LLM Prompts and Responses
Design a machine learning system that detects harmful content in prompts sent to, and responses produced by, a large language model. Tests policy-driven framing, labeling and adversarial data, layered classifiers under latency limits, per-category evaluation, streaming checks and monitoring.
ML System DesignOnsite - DatabricksMediumMachine Learning EngineerRetrieve and Rerank Tables and Notebooks for a Data-Aware Coding Agent
Design the retrieval system that lets an internal coding agent find the right tables and notebooks to answer data questions. Tests what to index for each asset type, hybrid retrieval plus reranking, handling noisy or stale assets, access control, and metric design from retrieval to final answers.
ML System DesignOnsite - DecagonMediumSoftware EngineerDesign a Single-Domain Chatbot
This question evaluates a candidate's ability to design production-grade conversational AI systems, testing competencies in retrieval-augmented generation, document ingestion and indexing, multi-turn session management, prompt construction, model serving, grounding and citation, safety, access control, monitoring, and operational trade-offs.
ML System DesignOnsite - DistylMediumMachine Learning EngineerDesign Streaming Speech and Multi-Model Chat Systems
A two-part system-design interview covering incremental multi-speaker transcription and a streaming multi-model chat service. Candidates must reason about ordering, revisions, backpressure, model fallback, durable conversation state, safety, and recovery.
ML System DesignOnsite - DoorDashMediumSoftware EngineerDesign a Store Recommendation System and Explain ML Trade-offs
Design a personalized store recommendation feed for an on-demand delivery platform and evaluate key machine learning modeling trade-offs.
ML System DesignOnsiteRecommendation SystemsMulti-Task LearningRanking - FaireMediumMachine Learning EngineerData quality, lineage, monitoring and backfill design for an ML feature store
A discussion-based ML infrastructure design round on keeping the data behind models reliable and well governed. It covers defining and measuring availability, freshness, completeness, validity, consistency, distribution, compliance and lineage, monitoring a feature store, and planning a safe, correct data backfill.
ML System DesignOnsite - FaireMediumMachine Learning EngineerDesign a real-time model inference platform: deployment, scaling and monitoring
An ML system design question asking for a real-time model inference platform that teams use to serve trained models inside their applications' request paths. It tests safe model deployment and rollback, per-model autoscaling and batching, failure handling, and monitoring from infrastructure up to model quality.
ML System DesignOnsite - FigmaMediumMachine Learning EngineerDesign Feed Ranking and HTML Generation
This problem tests the design of a dual-purpose ML system: a personalized content feed ranker and an LLM-based component that translates user intent into rendered HTML widgets.
ML System DesignOnsiteRankingLLMRecommendation Systems
Anthropic
Apple
Atlassian
C3 AI
Capital One
Cohere
Cresta
Databricks
Decagon
DoorDash
Faire
Figma