Cohere interview questions

Showing 1–14 of 14 results

• Answer Enterprise Questions Across Internal and External SourcesMedium

  Design enterprise AI question answering across internal and external sources with permission-aware retrieval, entity mapping, evidence-grounded synthesis, and freshness controls.

  CohereSoftware EngineerSystem DesignTechnical Screen
• Build a Credit Card Form With Validation and Linked Input FieldsMedium

  A front-end UI coding question: build a credit card form with card number, name, expiry and security code, where every field is validated and the fields are linked, so the detected card network changes formatting and security-code rules. It tests state design, caret-safe input formatting, error timing, accessibility and safe handling of card data.

  CohereFrontend EngineerSoftware Engineering FundamentalsOnsite
• Build an LLM-driven text-game agentMedium

  This question tests the ability to design a robust pipeline for an LLM-based agent interacting with a stateful environment, focusing on context management, output validation, and error recovery.

  CohereMachine Learning EngineerSoftware Engineering FundamentalsOnsite
• Defend Your Own Research Paper: Experimental Setup, Practicality and Later WorkMedium

  A research paper round where you present your own published paper to an interviewer who already knows it and questions its weaknesses. Prepare to judge whether the experimental setup was reasonable, whether the results are practical, and what the field has learned since publication, all with specific, honest self-critique.

  CohereMachine Learning EngineerMachine LearningOnsite
• Delete One '5' From a Digit String to Get the Largest NumberMedium

  Given a digit string containing at least one 5, delete exactly one 5 so the remaining string has the largest possible numeric value, keeping any leading zeros. This intern online-assessment problem tests reasoning about how each digit position affects a number and careful handling of repeated digits.

  CohereSoftware EngineerCoding & AlgorithmsOnsite
• Design How an LLM Detects Questions Beyond Its Training Cutoff and Says SoMedium

  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.

  CohereMachine Learning EngineerML System DesignOnsite
• Design Real-Time Comments with Optimistic UpdatesMedium

  Design optimistic real-time comments with temporary identities, idempotent posting, response/event reconciliation, visible failure states, and durable history.

  CohereFrontend EngineerSystem DesignOther
• Design a Post-Training Pipeline to Improve a 7B Model's Coding for an EnterpriseMedium

  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.

  CohereMachine Learning EngineerML System DesignOnsite
• Design a Real-Time Comment System Front End With Optimistic UpdatesMedium

  Front-end system design for a real-time comment feed, centered on optimistic updates: a new comment is appended to the list before the backend confirms it, so the user can keep typing. It tests client state modeling, reconciliation with live echoes, failure and retry handling, ordering, reconnection and rendering performance.

  CohereFrontend EngineerSystem DesignOnsite
• Explain LLM training, RL, and evaluationMedium

  This question tests your conceptual understanding of the end-to-end LLM lifecycle, specifically the transition from raw text pre-training to alignment and evaluation.

  CohereMachine Learning EngineerMachine LearningOnsite
• Explain Top-k and Top-p DecodingMedium

  This question tests your conceptual and mathematical understanding of Top-k and Top-p (nucleus) sampling, focusing on how they modify the output probability distribution of autoregressive language models to balance creativity and coherence.

  CohereMachine Learning EngineerMachine LearningOnsite
• Implement Autoregressive Decoding StrategiesMedium

  This problem tests your ability to implement common sequence decoding strategies used in large language models by manipulating token probability distributions.

  CohereMachine Learning EngineerCoding & AlgorithmsOnsite
• Improve LLM reasoning for a domain taskMedium

  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).

  CohereMachine Learning EngineerML System DesignOnsite
• Replace a Heuristic Terminal Text Game's Python Logic With LLM CallsMedium

  A live coding round: replace the heuristic Python implementation of a terminal text game with one driven by an LLM while keeping the game playable end to end. AI assistants are allowed, but the round judges your own implementation and debugging, including validating model output, fallbacks, and tests with a fake model.

  CohereMachine Learning EngineerSoftware Engineering FundamentalsOnsite