Machine Learning Engineer · Phone Screen · Multiple Types

Expedia — Machine Learning Engineer ❌ Failed

Level: Staff-Level

Round: Phone Screen · Type: Multiple Types · Difficulty: 6/10 · Duration: 60 min · Interviewer: Neutral

Topics: Reinforcement Learning, Large Language Models (LLMs), Search Engine Optimization (SEO), L1 Regularization, L2 Regularization, Graph Traversal, Spark, Data Analysis, Transformers, LSTMs, Neural Networks, He Initialization, SQL, Window Functions, Recommendation Systems, AWS Lambda, Feature Extraction, Data Pipelines

Location: Seattle, WA, US

Interview date: 2025-08-20

Got offer: False

Summary

This was a technical phone screen covering a range of machine learning and data engineering topics.

Details

During the phone screen & onsite interviews, I was asked about the following:

1. Using reinforcement learning to improve LLM performance for generating web content to improve SEO.
2. L1 and L2 regularization.
3. LeetCode problem: Reconstruct Itinerary.
4. LeetCode problem #198: Evaluate Division.
5. Using Spark to read a CSV file and analyze the data.
6. How much does processing time increase for Transformers vs. LSTMs if the amount of data doubles?
7. Probability question: 1% of the population has a disease. The diagnosis is 50% accurate. The diagnosis has a 1% false positive rate. If a person is diagnosed with the disease, what is the probability they actually have it?
8. Neural network He initialization: Why is there a square root, and why is the parameter 2?
9. SQL window function: Last seven days of sales, where each day is a row.
10. Designing a recommendation system (including the advantages and use cases of AWS Lambda).
11. Feature extraction and building data pipelines.

LeetCode similar: LeetCode Reconstruct Itinerary, LeetCode Evaluate Division