Accenture Interview Experience for LLM Operations Engineer

Last Updated : 27 Nov, 2025

Candidate Information:

  • Experience Level: 2 Years
  • Location: Remote/Bengaluru
  • Date: November 2025
  • Mode: Virtual Interview

I applied for the LLM Operations Engineer role at Accenture. My background is in Conversational AI, Oracle Digital Assistant (ODA), and Generative AI, with strong hands-on work building enterprise chatbots and RAG-powered systems for Oracle Fusion HCM.

The interview was technical and focused on:

  • Python + SQL basics
  • ML fundamentals
  • LLM + RAG concepts
  • ODA domain knowledge
  • Cloud cost awareness
  • JavaScript async fundamentals

Here is my complete interview experience, question-by-question.

Overview of Interview Process:

Technical Round

  • Duration : 1 hour

The interviewer began with introductions and then deep-dived into my projects, coding skills, and understanding of LLM operations.

Below is the exact set of questions I was asked.

  1. Tell me about yourself
  2. Explain your engineering project – problem statement and your role
  3. Explain your current project – problem statement and your role
  4. Write Python code to read a CSV and impute null values with median
  5. Write Python + SQLite query to find employees with more than 3 leave requests in the previous month
  6. How do you manage cloud costs?
  7. Explain prompting techniques
  8. Difference between regression and classification – name some algorithms
  9. What is overfitting?
  10. How would you deploy a multilingual chatbot in ODA?
  11. What is RAG and how did you use it in your project?
  12. Which metrics would you focus on to improve model performance?
  13. What kind of LLM models are available in the market?
  14. What are the data structures in Python?
  15. Difference between arrays and strings
  16. How would you ensure an intent in ODA gives the correct answer? What precautions would you take?
  17. How are async operations handled in JavaScript, and how does the event loop work?

Post-Interview Reflections:

  • I realized that I need to have to polish my concepts around Machine Learning also needed to work around my coding skills a bit.
  • I was unable to write both codes as well as unable to answer few questions. It is important to attempt them promptly, avoid cheating at all costs.
  • I fumbled while communicating, After promptly asking intrviewer about the feedback, he told me to work on my coding skills.
  • If selected, I was to work on cloud, preferably AWS or GCP.

Additional Information:

  • The next steps will depend upon the feedback by the interviewer. It might take an unspecified amount of time though.

Closing Note:

  • It was a balanced interview, part coding, part ML basics, part ODA understanding, and a major portion focused on LLM and RAG.
  • The interview was fair, and the feedback was helpful.
  • For now, I am continuing to prepare for the next rounds with stronger coding fundamentals.
Comment

Explore