Nvidia technical phone screen interview

Questions Nvidia asks in this round 27

• 

  Deep learning frameworks and compilers: from NumPy to PyTorch and JAXML designDiscuss where machine learning frameworks are heading: how they moved from NumPy to PyTorch and JAX, and how PyTorch and JAX differ (eager versus traced execution, functional…

  Possibly retiredWere you asked this too?
• 

  Key-value store with constant-time set-allCodingImplement a key-value store with set, get and a setAll operation that assigns one value to every key, all in O(1). A related variant manages dataset sizes under a fixed disk…

  Last asked Apr 2026Were you asked this too?
• 

  Infrastructure fundamentals: VMs, containers, Kubernetes and CI/CDCodingA senior infrastructure screen of conversation questions: 1. How do you bring up a new compute cluster in a new cloud account, and deploy Kubernetes there for the first time? 2…

  Last asked Jan 2026Were you asked this too?
• 

  Top K frequent elementsCodingGiven an integer array and k, return the k values that occur most often, in any order; ties at the cutoff may go either way. One version asks for the most frequent request IDs in…

  Last asked Jan 2026Were you asked this too?
• 

  Construct binary tree from inorder and postorder traversalCodingGiven the inorder and postorder traversals of a binary tree with distinct values, rebuild the tree. The last postorder value is the root; its position in the inorder list splits…

  Possibly retiredWere you asked this too?
• 

  Transformer and attention fundamentalsML designExplain how a transformer works from the inside: how self-attention is computed, why the scores are scaled, how encoder and decoder masking differ, and what multiple heads buy…

  Possibly retiredWere you asked this too?
• 

  Fetch paginated records and filter them by rules and a historical averageCodingAn HTTP endpoint serves patients' medical records, ten per page, with no page count given up front. Each record has many fields, most of them irrelevant to you. 1. Fetch every…

  Last asked Oct 2026Were you asked this too?
• 

  How do you use AI in your work?BehavioralDescribe how AI tools fit into your engineering work: where they help, where they fail, how you verify their output, and how they have changed the way you or your team work.

  Last asked Sep 2026Were you asked this too?
• 

  Quantize a float tensor to int8CodingConvert an FP32 tensor to INT8 with a given scale, rounding and clamping correctly. Then add asymmetric quantization with a zero point, and explain how you would choose scale and…

  Last asked Sep 2026Were you asked this too?
• 

  Count distinct strings with your own hash table, then on a GPUCodingGiven an array of strings, return how many distinct strings it contains, without using library hash sets or maps: build your own hash table with a hash function and collision…

  Last asked Jul 2026Were you asked this too?
• 

  Fixed-capacity circular bufferCodingImplement a generic FIFO buffer with a fixed capacity, allocated once up front, with constant-time push and pop. Discuss how to signal full and empty and how the buffer would be…

  Last asked Jul 2026Were you asked this too?
• 

  Firmware fundamentals: endianness, ring buffers and the stackCodingAnswer and code short firmware questions: what big- and little-endian mean and how to detect or convert byte order, how a ring buffer works between an interrupt handler and the…

  Last asked Jul 2026Were you asked this too?
• 

  Design a control plane that tracks hosts and assigns jobsSystem designDesign a central service that knows the live state of many compute hosts and hands jobs to them. Focus on how hosts and the service talk (heartbeats, push versus pull), where…

  Last asked Jun 2026Were you asked this too?
• 

  Merge intervalsCodingGiven a list of intervals (for example time ranges), merge every group of overlapping intervals and return the result in sorted order.

  Last asked May 2026Were you asked this too?
• 

  List every file path in a directory treeCodingGiven a directory tree (folders containing files and other folders), return the full path of every file. Write it with a depth-first search, then discuss an iterative version and…

  Last asked Mar 2026Were you asked this too?
• 

  LRU cache with runtime resizingCodingImplement a least recently used (LRU) cache whose capacity can change while it is in use. - get(key) returns the value for key, or -1 if it is not cached. A successful get makes…

  Last asked Feb 2026Were you asked this too?
• 

  Polynomial multiplication API in CCodingDesign and implement a small C API for polynomials: a representation, creation and freeing, and a function that multiplies two polynomials and returns a new one. Pay attention to…

  Last asked Feb 2026Were you asked this too?
• 

  Data structure fundamentals: arrays, linked lists, hash tables and treesCodingAnswer quick fundamentals questions: the time complexity of common sorting algorithms; arrays versus linked lists, including the cost of inserting at the head and in the middle…

  Last asked Feb 2026Were you asked this too?
• 

  Find the nth primeCodingReturn the nth prime number. Start with trial division, then use a sieve of Eratosthenes with an estimated upper bound and explain the complexity of both.

  Last asked Feb 2026Were you asked this too?
• 

  Total population of a city across zip codes in SQLCodingFour tables describe geography: country, state, city and zip code, each pointing to its parent, and the population is stored per zip code. Write a query that returns the total…

  Last asked Feb 2026Were you asked this too?
• 

  String compression and decompressionCodingWrite a function that compresses a string (for example with run-length encoding, aaabcc to a3b1c2 or a variant the interviewer gives) and one that decompresses it back exactly…

  Last asked Jan 2026Were you asked this too?
• 

  Bridge and torch problemCodingPeople must cross a bridge at night with one flashlight. At most two cross at once, a pair moves at the slower person's speed, and someone has to bring the flashlight back. Given…

  Last asked Jan 2026Were you asked this too?
• 

  MLP tensor shapes and PyTorch broadcastingML designGiven a small multi-layer perceptron and a batch of inputs, write down the shape of every weight, bias, activation and gradient. Then explain PyTorch's broadcasting rules, how the…

  Last asked Jan 2026Were you asked this too?
• 

  Group anagramsCodingGroup the words in a list that are anagrams of one another, keeping repeated words. Build the grouping key from letter counts instead of sorting each word.

  Last asked Dec 2025Were you asked this too?
• 

  Simple VM manager with add, list, modify and deleteCodingWrite a simple manager for virtual machines: add a VM with its attributes, list VMs, modify a VM's attributes and delete a VM. Handle errors such as adding a duplicate id or…

  Last asked Nov 2025Were you asked this too?
• 

  Design a distributed counter for many usersSystem designMany users increment shared counters concurrently. Start with a correct single-process counter with locking, then design it for many servers: atomic increments in a database…

  Possibly retiredWere you asked this too?
• 

  Design an artifact store on Kubernetes and CassandraSystem designA Java web API runs in Kubernetes on top of a Cassandra cluster and stores artifacts, each identified by a user-given name. An artifact must be created only once, even when…

  Possibly retiredWere you asked this too?