AI Software Engineer

Description

Interview questions for AI Software Engineer

Interview Questions (36)

Q1What is a RAG pipeline and how to design it using Python code?
Technical
Q2How do you deal with hallucination in LLMs?
Technical
Q3How are OpenAI embeddings different from normal deep learning embeddings?
Technical
Q4What is semantic searching?
Technical
Q5What are your metrics for evaluating RAG performance?
Technical
Q6Talk about prompt engineering principles.
Technical
Q7Explain LangGraph / LangChain.
Technical
Q8How would you optimize a RAG system that is currently returning noise, which increases context entropy and causes attention decay in the token sequence?
Technical
Q9Explain precision/recall tradeoff.
Technical
Q10What is F1 score?
Technical
Q11What are underfitting and overfitting, and the tradeoff of each?
Technical
Q12Explain the significance of the ROC (Receiver Operating Characteristic) curve in machine learning.
Technical
Q13Choose Linear Regression or Logistic Regression. Explain your pick and how you implement it.
Technical
Q14Explain Gradient Descent in machine learning.
Technical
Q15What algorithms would you use for a classification problem?
Technical
Q16What is your favourite ML model?
Experience
Q17Can you explain the basic principles behind Generative Adversarial Networks (GANs)?
Technical
Q18Can you explain the architecture of a convolutional neural network (CNN)?
Technical
Q19Tell me about BERT (Bidirectional Encoder Representations from Transformers) architecture.
Technical
Q20Explain RNN (Recurrent Neural Network) Transfer Learning.
Technical
Q21Explain a project where you implemented a machine learning model.
Experience
Q22How do you deal with certain situations in model training?
Situational
Q23Discuss a machine learning model you have experience with and its underlying principles.
Experience
Q24What is fine-tuning? How would you fine-tune X company’s model which uses hosted LLM models?
Technical
Q25Describe any AI pipelines that you implemented in any cloud.
Experience
Q26What does it mean for an algorithm to be numerically unstable?
Technical
Q27What are the trade-offs between using FP32, FP16, and BF16 precision when training large-scale models on NVIDIA GPUs?
Technical
Q28When dealing with massive datasets, how do you decide between using sparse matrix representations versus dense ones, and how does this affect memory bandwidth?
Situational
Q29Precision vs Recall — which matters more for fraud detection?
Situational
Q30How would you integrate an AI-based system to help a zoo owner improve their business?
Situational
Q31Design a recommender system.
Situational
Q32Design and implement a full system solution.
Situational
Q33Design a chatbot.
Situational
Q34Python: immutable vs mutable variables passed to a function — how different are they?
Technical
Q35Python: what is the difference between is and ==?
Technical
Q36How would you build reproducible code?
Technical