Accenture 2nd round interview questions 1. What is more comfortable to you ? ML/DL or genAI? 2. Any exposure to genAI, have you done any POC? 3. Have you done any course or genAI? 4. In Deep learning what kind of pretrained models you are using? 5. Which sequence to sequence you have used? 6. What sort of architecture sequence to sequence model has? (Interviewer answer is encoder and decoder model) 7. For deep learning which library you used? 8. Any other framework do you aware that deep learning models can be trained? 9. What you think most difficult part while you implement sequence to sequence model in tensorflow? 10. Give a objective of a deep learning model like this is the business objective, due to which I want to implement sequence to sequence architecture? how prepared data for the same?then how you implement sequence to sequence modelling? 11. Let’s assume your model is giving high False Negative that you don’t want and you need to improve this model, what performance tuning you applied in the modelling? 12. Let’s assume you done all fine tuning but still you are not achieving you model performance then what else you can do? Don’t you think synthesizing data would be good option? 13. What are the hyper parameters you tuned in deep learning model? 14. Adding a deep layer would suffice your problem do you think? Is it have any adverse effects? 15. In Random Forest what is out of bag error? 16. What is precision medicine or (might be precision magascin)? as you are working in Pharma domain so long. 17. In pharma field what are the advance use cases you heard of as for as AI is concern? 18. What is the dimension of the numpy array? 19. If you find programmatically, how would you find the dimension of each of the provided numpy arrays? 20. What is the use of reshape in numpy? 21. How many use cases did you solve with DL & ML? 22. How many do done personal projects? 23. Which cloud flatform are you using? 24. In AWS what you have used to deploy models? 25. What is well known platform in aws to deploy models?