Recently, I had the opportunity to interview with **Quantrium** for an **Internship + PPO**, and I’m happy to share that I was selected 🎉. Overall, the interview process was well-structured, insightful, and focused on real-world understanding rather than rote learning. Below is a detailed breakdown of my experience. ## Interview Process Overview The selection process consisted of **four rounds**: 1. Resume Shortlisting 2. Online Assessment (OA) 3. Technical Round 1 (Senior Engineer) 4. Technical Round 2 (CTO & Project Manager) ## 1. Resume Shortlisting This was the initial screening round. Only candidates with **hands-on projects in the AI/ML domain** were shortlisted. Having relevant, well-documented projects played a crucial role here ## 2. Online Assessment (OA) The OA had **two sections**: * **MCQs** on AI/ML fundamentals * **Two coding problems** of medium difficulty (LeetCode-level) The coding questions tested logical thinking and problem-solving rather than obscure tricks. ## 3. Technical Round 1 (Senior Engineer) This round lasted about **1.5 hours** and was highly technical. ### Resume & Project Discussion It started with an introduction, followed by an in-depth discussion of my resume and projects. The interviewer focused on: * Why I chose specific tools/technologies * Trade-offs between different approaches * How design choices would impact scalability and performance **Tip:** Be honest, and know your projects deeply. You should be able to justify every design decision. ### AI/ML, DL, NLP & RAG The discussion then moved to AI/ML fundamentals, Deep Learning, NLP, and **RAG-based applications**. I was asked to: * Design a RAG system * Explain each pipeline component in detail * Diagnose issues like hallucination even when retrieval is correct * Discuss document fusion, query transformation, and other RAG challenges In ML, I was given a **problem statement** and asked which model I would choose (e.g., Logistic Regression vs Random Forest). The interviewer went deep into Random Forest to understand my **thinking process and model selection rationale**. ### OOPs & DSA In the final part: * OOP concepts in Python were discussed, and I wrote code * Two DSA problems were asked: * Two Sum * Kth Largest Element in an Array I was allowed to code in either Python or C++. The round ended with a Q&A, where I asked about the company’s tech stack, work culture, and interview feedback. ## 4. Technical Round 2 (CTO & Project Manager) This round was conducted **the very next day**. ### CTO Round (System Design & R&D Focus) The CTO focused more on **production readiness and system design** rather than implementation details. Questions included: * How to scale my project as users increase * My R&D experience and tools used * Model performance over time * Fine-tuning techniques and approaches * Ensuring correctness and reliability in RAG-based systems He also asked about **Explainable AI**. While I wasn’t deeply familiar with it, I answered based on intuition and logical reasoning, which he appreciated. ### Project Manager Round (Process & Behavioral) The Project Manager focused on teamwork and execution: * My experience working in groups * Project management approach (**Agile**) * Differences between Agile and Waterfall * Behavioral questions like: * Why Quantrium? * How I handle conflicting ideas within a team These were straightforward if you’ve worked in collaborative environments. At the end, I asked about: * Company culture * Growth trajectory * How I could contribute to the company’s growth in the next 6 months ## Final Thoughts The entire process tested **depth of knowledge, clarity of thought, system-level understanding, and teamwork skills**. It was less about memorization and more about how you approach real-world problems. I’m grateful for the experience and excited about what lies ahead at Quantrium 🚀.

## I am sharing my interview experience with ICICI from last year -2025. I applied it through my campus placement drive as a fresher. ### ICICI came to IIT Delhi's placement season 2024–2025. It was listed as an ICICI Manager I. Manager I is a grade. It was open for all departments with no minimum CGPA criteria. # My profile at that time: I had an intern at a startup where I had built a recommendation model. I had been grinding LeetCode and Codeforces along with machine learning. I had successfully completed Andrew Ng's specialization course on ML and deep learning and practiced on Kaggle. # Online Assessment — 60 minutes At first we have CV shortlisting, where I got shortlisted. After that, I have to give an online assessment that contains only MCQ questions related to OOPs, DSA, Cpp, Java, and Python, which I successfully cleared. After that, I got a call from the placement cell that I have to give a personality test called personality profiler, where they asked me about how I will behave or react under some given conditions. Update 1—The interview shortlist comes on 30th November around 5 pm. Update 2—The interview was scheduled on the morning of 3rd December # Interview: It was 9 in the morning; the atmosphere was quite foggy and cold, and I was quite nervous. The interview is going to be the only one round with both technical and HR at the same time. ## They called me for Interview First question: “Tell about yourself.” I had prepared for this question and told about my name, my hometown, and my specializations along with my hobbies. Second question: “Explain any one of your projects.” I know this question will pop up, and I had prepared for this question by preparing my best project in depth. But as ICICI is a bank, I thought I should explain my project, which is relevant to the bank in any way. So instead of explaining about the project ‘AI Interviewer,’ I choose stock analysis by AI agent. I had been asked some cross questions regarding my project, which I answered successfully. Third Question: “Can you optimize the time taken to train ANN?” As there are many ways to optimize artificial neural networks. It starts with the hardware side, like using more powerful GPUs. Then I explained we can drop neurons in order to train the model faster. Then he asked me about some mathematical approaches, so I recalled Adam optimization, and I explained it in detail by drawing some formulas and images. Fourth question: “Forward and Back Propagation.” I started with forward propagation by considering a 2-layer neural network and similarly derived back propagation from it, which he seems satisfied with. ## Update 3 — Now my technical round is over and its for HR round 5. She asked me about why ICICI, my background, my future goals, and general HR questions, which I prepared before the interview and answered optimally, at least according to me. Update 4: I got a call from POC that they were willing to offer me, which I gladly accepted. I felt a surge of excitement and joy. Update 5: HR called me and congratulated me, after which we had a professional handshake and goodbye. 3rd December 2024 will be a special day for me as I get my first job. It was my first and last interview at IIT Delhi. Thanks for reading.

There is a techinical round and interview round happend in the same day(offline mode) where we were given an hour or so to write the test of MCQs which were mostly theoretical and some aptitudes. The level of the paper is moderate where the domains of questions were analog, communication, transmition lines, antennas and Signal processing. The imediately corrected the paper and given us the result for next round after half an hour. The interview rounds were friendly some comm. question like amplitude and frequency modulations, project explaination, Which modulation is less noisy all stuff. Then question of hobbies and general questions were followed.