Subex Data Scientist Interview Questions and Answers

Q1. What is Overfitting in decision tree
Ans. 
Overfitting in decision trees occurs when the model learns noise in the training data rather than the underlying pattern.
• Overfitting happens when the decision tree is too complex and captures noise in the training data.

• It leads to poor generalization on unseen data, as the model is too specific to the training set.

• To prevent overfitting, techniques like pruning, setting a minimum number of samples per leaf, or using en
Answered by AI
Add your answer
Q2. What is bagging
Ans. 
Bagging is a machine learning ensemble technique where multiple models are trained on different subsets of the training data and their predictions are combined.
• Bagging stands for Bootstrap Aggregating.

• It helps reduce overfitting by combining the predictions of multiple models.

• Random Forest is a popular algorithm that uses bagging by training multiple decision trees on random subsets of the data.
Answered by AI
Add your answer
Q3. What is neuron How its used in DL
Ans. 
A neuron is a basic unit of a neural network that receives input, processes it, and produces an output.
• Neurons are inspired by biological neurons in the human brain.

• They receive input signals, apply weights to them, sum them up, and pass the result through an activation function.

• Neurons are organized in layers in a neural network, with each layer performing specific tasks.

• In deep learning, multiple layers of neurons ar...read more
Answered by AI
Add your answer

Basic python sql questions, mcq based and coding questions.

Q1. They asked the projects which I am working on and had discussion on the same. Asked few statistical questions like boxplots and normal distribution. Also asked basic questions on advance SQL
Add your answer

Q1. Problem solving questions. Gave a case study to check presence of mind
Add your answer

Q1. Basic salary discussion
Add your answer

15 MCQ, 2 coding round

Q1. What is PCA explain eigen values
Add your answer
Q2. How chatbot works really
Add your answer

Python sql basic questions

Q1. Explain Transformers how different from previous RNN, LSTM etc.
Ans. 
Transformers are a type of neural network architecture that utilizes self-attention mechanisms to process sequential data.
• Transformers use self-attention mechanisms to weigh the importance of different input elements, allowing for parallel processing of sequences.

• Unlike RNNs and LSTMs, Transformers do not rely on sequential processing, making them more efficient for long-range dependencies.

• Transformers have been shown ...read more
Answered by AI
Add your answer
Q2. What are different types of Attention?
Ans. 
Different types of Attention include self-attention, global attention, and local attention.
• Self-attention focuses on relationships within the input sequence itself.

• Global attention considers the entire input sequence when making predictions.

• Local attention only attends to a subset of the input sequence at a time.

• Examples include Transformer's self-attention mechanism, Bahdanau attention, and Luong attention.
Answered by AI
Add your answer
Q3. Difference between GPT and BERT model
Ans. 
GPT is a generative model while BERT is a transformer model for natural language processing.
• GPT is a generative model that predicts the next word in a sentence based on previous words.

• BERT is a transformer model that considers the context of a word by looking at the entire sentence.

• GPT is unidirectional, while BERT is bidirectional.

• GPT is better for text generation tasks, while BERT is better for understanding the cont
Answered by AI
Add your answer

Q1. Difference between Data scientist, ML and AI
Ans. 
Data scientists analyze data to gain insights, machine learning (ML) involves algorithms that improve automatically through experience, and artificial intelligence (AI) refers to machines mimicking human cognitive functions.
• Data scientists analyze large amounts of data to uncover patterns and insights.

• Machine learning involves developing algorithms that improve automatically through experience.

• Artificial intelligence r...read more
Answered by AI
Add your answer

Python coding and ML resume based questions

Q1. Past Projects with the Director
Add your answer
Q2. Behavioral Round questions Mostly
Add your answer

First round is coding round where two use cases are there. Need to solve them

Q1. They will all topics Statistics, SQL, Python, Machine Learning, Data Science
Add your answer

Q1. They will discuss more on the projects what we worked on
Add your answer

Q1. Salary Discussion
Add your answer

Q1. Questions on Transformer Architecture and
Add your answer
Q2. System design tradeoffs and basic principles
Ans. 
System design tradeoffs involve balancing various factors to optimize performance and efficiency.
• Consider scalability, reliability, latency, and cost when designing systems

• Tradeoffs may involve sacrificing one aspect for the benefit of another

• Examples include choosing between consistency and availability in distributed systems
Answered by AI
Add your answer

Q1. Various questions on my projects
Add your answer
Q2. NLP based questions and metrics calculation and case study
Add your answer

Q1. Basic HR questions
Add your answer
Q2. Why Fractal, etc
Ans. 
Fractals are used in data science for analyzing complex and self-similar patterns.
• Fractals are useful for analyzing data with repeating patterns at different scales.

• They are used in image compression, signal processing, and financial market analysis.

• Fractal analysis can help in understanding the underlying structure of data and making predictions.
Answered by AI
Add your answer