ZS LLM Engineer Interview Experience

Last Updated : 11 Dec, 2024

I was approached for the LLM Engineer role at ZS through a third-party agency that found my profile on Naukri.com. After expressing interest, I submitted the required details for the application.

Background

  • Previous Experience: 14 months as an AI Engineer at a startup.
  • The selection process involved three rounds:

Round 1: Resume Shortlisting

My resume was shortlisted, and ZS’s HR reached out to schedule the next steps. They expressed interest in my profile, finding it a promising fit for the role.

Round 2: Technical Interview

Duration: 1 hour

This round began with a brief introduction and discussions on my previous work experience and projects. The questions covered a mix of technical and coding topics:

Key Topics Covered

  1. RAG Application Development:
    • Procedure to build a notebook-based RAG (Retrieval-Augmented Generation) application.
    • Concepts like VectorBases and steps involved in RAG.
  2. Coding Challenges:
    • Find the longest repeated substring that appears at least twice in a string (e.g., for s='banana', the result is 'ana').
    • Detect a loop/cycle in a linked list.
  3. LLM Benchmarking:
    • Definitions of MMLU, HELM, HumanEval, and Big-Bench.
  4. Hyperparameters in LLMs:
    • Concepts of Top-K, Top-P, Temperature, etc.
  5. Advanced Concepts:
    • CoT (Chain of Thought) reasoning.
    • Self-Attention model, Encoder, Decoder, and their roles.
    • Generator and Discriminator used in GANs.

Towards the end of the interview, I sought feedback on my performance and advice for improving my AI skills.

Round 3: Experience-Based Interview

Duration: 30-45 minutes

This round was conducted by a senior LLM engineer or a ZS partner. It was largely based on my resume and work experience.

Key Questions Asked

  1. Introduction: Briefly introduce yourself.
  2. About ZS:
    • What do you know about ZS?
    • Why do you want to join ZS?
  3. Work Experience:
    • Detailed discussion on the GenAI products I built, including their usage metrics.
  4. Personal Questions:
    • Family background.
  5. AWS Experience:
    • Questions about the AWS tools mentioned in my resume.
    • Specific discussion on my experience with AWS Bedrock.
  6. Latest Advancements in GenAI:
    • Insights on topics like Hybrid RAG, multimodal agents, and MoE LLAVA.

Closing Questions

At the end, I asked about:

  • The team structure at ZS.
  • New products they are developing.
  • The interviewer’s experience of being part of ZS.

Final Thoughts

The interview process was highly engaging and covered a mix of technical depth and real-world applications of LLMs and GenAI. It was a great opportunity to reflect on my skills and learn about the innovative projects at ZS.

Comment

Explore