- Briefly tell us about yourself and one of your most interesting projects.
- Implement a class or function that, given a corpus of strings, computes TF-IDF.
- How would you store and optimize TF-IDF with 10 thousand documents and a vocabulary of several million words? What sparse matrix formats exist?
- What are the time and space complexity of fit and transform in TF-IDF?
- What is the GIL, and is there parallelism in Python?
- There are two dense NumPy matrices of 2 GB each; they need to be multiplied on a machine with 8 GB of RAM. Will the computation fit in memory, and how will it be executed?
- What is the difference between asynchrony and multithreading, and when is each better to use?
- What is a decorator in Python?
- Write a try/retry decorator that will repeatedly execute any function until it completes successfully.
- How does Word2Vec work, how is it trained, and what is fed as input? What architecture does it have?
- In Word2Vec, will the word "castle" have the same vector in the sentences "I opened the castle" and "I went into the castle"?
- How can the problem of homonymy in static embeddings before attention and BERT be solved?
- What is GloVe and what is its essence?
- How does self-attention work, and what is its asymptotic complexity in time and memory?
- What is FlashAttention, how is it better than ordinary attention, and does it improve time, memory, or both parameters?
- What is PagedAttention, and how does it optimize memory usage?
- How does a transformer take word order into account, and how is RoPE applied in code?
- What is the difference between a bi-encoder and a cross-encoder, and when should each be used?
- What is LoRA and why is it needed?
- Calculate the number of parameters of a LoRA adapter for a 500x1000 matrix with rank=8.
- How would you estimate the size of a LoRA adapter for a model with approximately 1 billion parameters at rank=8?
- What types of search are used in RAG, and what are the advantages and disadvantages of each?
Raiffeisenbank Machine Learning Engineer Interview Experience — 22 Questions from TF-IDF to RAG
Technical Screenhard
Published
Curated and edited by PracHub
Practice the questions from this interview
Self-Introduction and Walkthrough of Your Most Interesting Project for an AI Engineer RoleBehavioral & Leadership
TF-IDF From Scratch: Implementation, Sparse Storage for Millions of Terms, ComplexityMachine Learning
Python GIL, Async vs Threads, and Multiplying Two 2 GB NumPy Matrices in 8 GB RAMSoftware Engineering Fundamentals
Python Decorators and a Retry Decorator That Re-Runs a Function Until It SucceedsSoftware Engineering Fundamentals
Word2Vec Training, Homonymy in Static Embeddings, and the Idea Behind GloVeMachine Learning
Self-Attention Complexity, FlashAttention, PagedAttention, and RoPE in CodeMachine Learning
Bi-Encoders vs Cross-Encoders and the Search Methods Used in RAGMachine Learning
LoRA: Why It Is Needed and How to Count Adapter Parameters at Rank 8Machine Learning
Discussion
Loading comments…