Comcast India Engineering Center Data Scientist Interview Questions and Answers

Q1. Basicn details to check for qualifications
Add your answer

Q1. About my projects
Add your answer

Q1. More details about ML models
Add your answer

Python Programming related questions, along with one advanced SQL query problem. The final question was a Data Science project on prediting sales potential of various outlets.

Q1. Pattern based - Three memory chips, each of 1GB. You have to store 3GB of data in these chips in such a way that even if one memory chip is corrupted, no data is lost.
Ans. 
Use RAID 5 to store data across all three memory chips with parity bits for fault tolerance.
• Implement RAID 5 to distribute data and parity bits across all three memory chips.

• If one memory chip is corrupted, the data can be reconstructed using the parity bits from the other two chips.

• Example: Store 1GB of data on each chip and use the remaining space for parity bits to ensure fault tolerance.
Answered by AI
Add your answer
Q2. DSA question - Get the longest common prefix string from a list of strings
Ans. 
Find the longest common prefix string from a list of strings.
• Iterate through the characters of the first string and compare with corresponding characters of other strings

• Stop when a mismatch is found or when reaching the end of any string

• Return the prefix found so far
Answered by AI
Add your answer
Q3. DBMS question - What are joins and what are their types?
Ans. 
Joins are used in DBMS to combine rows from two or more tables based on a related column between them.
• Types of joins include INNER JOIN, LEFT JOIN, RIGHT JOIN, and FULL JOIN.

• INNER JOIN returns rows when there is at least one match in both tables.

• LEFT JOIN returns all rows from the left table and the matched rows from the right table.

• RIGHT JOIN returns all rows from the right table and the matched rows from the left tab...read more
Answered by AI
Add your answer
Q4. ML and DL question - Activation functions, Exploding and Vanishing Gradient problems, LSTM and GRU models frameworks and differences.
Add your answer

Was taken by the product manager employed in the company. Basic case study question regarding a ride share app planning to expand internationally.

A formal orientation and introduction with the VP and founder of ION India

Q1. Introduce yourself and mention your work that makes you relevant for the job.
Add your answer
Q2. Is there any correlation between algorithms and law?
Ans. 
Algorithms and law can be correlated through the use of algorithms in legal processes and decision-making.
• Algorithms can be used in legal research to analyze large amounts of data and identify patterns or trends.

• Predictive algorithms can be used in legal cases to assess the likelihood of success or failure.

• Algorithmic tools can help in legal document review and contract analysis.

• However, there are concerns about bias i...read more
Answered by AI
Add your answer

Q1. Probability related
Add your answer
Q2. Stats related like IQR etc.
Add your answer

Q1. Difference between entropy & information gain
Ans. 
Entropy measures randomness in data, while information gain measures the reduction in uncertainty after splitting data.
• Entropy is used in decision trees to measure impurity in a dataset before splitting it.

• Information gain is used in decision trees to measure the effectiveness of a split in reducing uncertainty.

• Entropy ranges from 0 (pure dataset) to 1 (completely impure dataset).

• Information gain is calculated as the d...read more
Answered by AI
Add your answer
Q2. LSTM & GRU, Which to use when ?
Ans. 
LSTM for longer sequences, GRU for faster training and less complex models.
• Use LSTM for tasks requiring long-term dependencies and memory retention.

• Use GRU for faster training and simpler models with fewer parameters.

• Consider using LSTM for tasks like language translation or speech recognition.

• Consider using GRU for tasks like sentiment analysis or text generation.
Answered by AI
Add your answer

Time Series data were given, we have to provide some insights

It contain both Aptitude and Coding about base models and Deep learning too

Q1. What different models technique ?
Ans. 
Different models techniques include linear regression, decision trees, random forests, support vector machines, and neural networks.
• Linear regression is used for predicting continuous values.

• Decision trees are used for classification and regression tasks.

• Random forests are an ensemble method based on decision trees.

• Support vector machines are used for classification tasks.

• Neural networks are used for complex pattern re
Answered by AI
Add your answer
Q2. What are performance metric where to use what?
Ans. 
Different performance metrics are used for different types of machine learning models to evaluate their effectiveness.
• For classification models, metrics like accuracy, precision, recall, F1 score, and ROC-AUC are commonly used.

• For regression models, metrics like mean squared error (MSE), mean absolute error (MAE), and R-squared are commonly used.

• For clustering models, metrics like silhouette score and Davies-Bouldin in...read more
Answered by AI
Add your answer

Q1. Explain about Project
Add your answer
Q2. What are problems faced in that project?
Add your answer

Q1. About projects and then questions related to ML and DL. Mostly focused on DL part
Add your answer
Q2. What is the difference between Adam optimizer and Gradient Descent Optimizer?
Ans. 
Adam optimizer is an extension to the Gradient Descent optimizer with adaptive learning rates and momentum.
• Adam optimizer combines the benefits of both AdaGrad and RMSProp optimizers.

• Adam optimizer uses adaptive learning rates for each parameter.

• Gradient Descent optimizer has a fixed learning rate for all parameters.

• Adam optimizer includes momentum to speed up convergence.

• Gradient Descent optimizer updates parameters b...read more
Answered by AI
Add your answer
Q3. When to use Relu and when not?
Ans. 
Use ReLU for hidden layers in deep neural networks, avoid for output layers.
• ReLU is commonly used in hidden layers to introduce non-linearity and speed up convergence.

• Avoid using ReLU in output layers for regression tasks as it can lead to vanishing gradients.

• Consider using Leaky ReLU or Sigmoid for output layers depending on the task.

• ReLU is computationally efficient and helps in preventing the vanishing gradient prob...read more
Answered by AI
Add your answer

Q1. Asked about projects. What is classification? Is knn used for regression?how? decision tree working for regression and classification Is naive Bayes used for regression?how? LLM Docker Aws GenAI Code for ...read more
Add your answer

Sql and python questions were there with basic logic check

Q1. Python code with funticion
Ans. 
Python code with function
• Define a function using 'def' keyword

• Include parameters inside parentheses

• Use 'return' statement to return a value from the function
Answered by AI
Add your answer
Q2. Sql with case statwment
Add your answer