Equations for: SGD, Momentum, L1 and L2 loss, linear regression etc. Statistics deep dive: maximum likelihood estimates, hypothesis testing, PDF etc. Deep Learning: number of parameters in a model, coding Conv, MaxPool and embedding layer etc
Applied Scientist Interview Questions
1,167 applied scientist interview questions shared by candidates
one of the coding question is lc46
Talk about your relevant experience
1. How would you design a pipeline for a dataset comprises students, class and scores
- Tell me why you're interested in this domain. - Tell me about relevant projects you've done in this domain. - What implications would a multi-modal generative AI tool have (compared to a unimodal one) and how would you go about mitigating its biases. - Case Study: A company wants to use an AI tool to summarize managers' feedback for employees. How would you "systematically" approach identifying the biases of such a tool?
Leetcode + LLM knowledge first a quick introduction from both sides then one medium level leetcode question followed by a lot of LLM related questions, also asked about projects on resume
Problem: Top K Most Similar Documents You are given: an integer array queryEmb of length D, representing a query embedding a 2D integer array docEmbs of size N x D, representing N document embeddings an integer k All embeddings are already L2-normalized. The cosine similarity between two normalized vectors is equal to their dot product. Return the indices of the k documents with the highest cosine similarity to queryEmb, ordered from most similar to least similar. If k > N, return all document indices sorted by similarity. Function Signature def topKSimilar(queryEmb: np.ndarray, docEmbs: np.ndarray, k: int) -> np.ndarray
How can you convince me that a number is irrational?
1, Self Introduction 2. Coding questions related to Linear regression, and basic ML
In which situations would you use different types of machine learning models, and what are some basic ideas behind transformers?
Viewing 941 - 950 interview questions