SmartSense Consulting Solutions

1. 1

   Online test

   MCQ test
   CodeSignal75 min60 questions

   machine learning · deep learning · CNN · NLP · computer vision · DBMS · SQL · OS · DSA · LLM agents and RAG

   Mixed single and multiple choice; little time per question.

2. 2

   Take-home assignment

   Case study: build agent plus RAG chatbot
   hard

   RAG · agents · SQL and vector database · report generation · deployment

   Two days given; data and specifications supplied; submit repository, video and presentation. Candidate called it very challenging.

3. 3

   Technical interview

   Interview round 1
   Classroom120 minoffline

   OS · DBMS · algorithms · scenario-based design · resume projects · Git · LLM architecture · positional encoding · agents · RAG · SQL · Python · ML · DL

   Each resume project questioned one by one.

4. 4

   Technical interview

   Interview round 2
   Classroom120 minoffline

   DSA · SQL queries · AWS · Docker · LangChain · FastAPI · DBMS scenarios · product deployment · CNN · LLMs · RAG · agents · resume projects

   Same day as round 1.

5. 5

   HR interview

   HR / behavioral
   90 mineasy

   about yourself · gaps in earlier rounds · puzzle

   Described as lighter than the two technical rounds.

Advice from candidates

38 pieces of advice, in their own words. Names removed.

• You do not need to produce a full-fledged application. The recruiters want to understand your attempt and approach to solving a problem.

  Speech-to-text task · 2024-25

• It is an AI/ML Engineer I role so you do not need to give your all efforts to task, just do the basic things and keep yourself ready for interview next day instead of keeping yourself awake the entire night and sleeping in the early morning interview.

  Speech-to-text task · 2024-25

• Revise Machine Learning, Deep Learning, NLP very well. There were some questions on Computer Vision as well.

  test · 2024-25

• If you revise your ML, DL, NLP, CV courses well, the interview round will be easy,

  interview · 2024-25

Loading the other 34…

Question patterns reported

The kind of problem, not the problem itself. Our practice questions are written to match these patterns.

• Pre-placementbuild a multi-agent AI system that summarises and classifies, logging queries and replies in a database
• Onlinememory management types in LLM agents
• Take-homebuild deployed chatbot from scratch in two days with demo video and slides
• Technicalwrite a Python program in an editor during the interview
• Technicalone DSA problem and two SQL query-writing tasks
• Take-homedesign an AI service platform in a deck of at most eight slides