Ik heb via een recruiter gesolliciteerd. De procedure duurde 4 weken. Ik heb gesolliciteerd bij Pattern in jul 2025
Sollicitatiegesprek
First round was a general round checking python and SQL skills. The questions were apt to what generally encompasses data science domain and not random DSA questions. The interviewer then asked questions from resume checking fundamentals and verifying if you have actually worked on things you mention in resume.
2nd round was a very thorough discussion of 1.5 hours with a very vague case study. One is supposed to define the problem statement, data, modelling process, evaluation techniques etc. During this, the interviewer went into depth in a lot of ML modelling techniques and overall approach of solutioning.
Sollicitatievragen [2]
Vraag 1
Round 1:
The questions were SQL join conceptual and small implementation question followed by pandas questions to manipulate dataframe and calculate rolling average of lags.
Round 2:
Many questions and cross questions on your approach to the case study - why would you choose this data, what else could you do. Other what-if scenarios on this. In modelling, fundamental questions of how some model works - not superficial level answers but demanded thorough answers.
It doesn’t feel like they are actually hiring. It seems they just want to show on LinkedIn that they are hiring. One of my contacts had a similar experience. Both of us appeared for interviews, and the interviews went very well, but we were told that they hired another person. Then, the same vacancy was reposted on LinkedIn the very next day. Maybe they don’t even have the budget in the first place. They were trying to hire senior-level candidates within a salary range of 10–15 LPA.
Sollicitatievragen [1]
Vraag 1
Normal around resume ml stuff doesn't seem anything different
Ik heb online gesolliciteerd. De procedure duurde 3 weken. Ik heb gesolliciteerd bij Pattern (Poona) in mei 2025
Sollicitatiegesprek
I initiated the process by contacting an HR directly on LinkedIn for a referral. This was followed by a brief screening call where he asked about 10 basic questions before scheduling the first formal interview. And lots of unncessary background things like college name, uni name grades etc.
Round 1: basic coding questions focusing on Pandas and SQL.
Round 2: Case Study for the team applied - It was on demand forecasting for multiple inventory items. We had a detailed discussion on relevant approaches, including using XGBoost, various feature engineering techniques, and model selection.
Round 3: VP Discussion - This was entirely a resume deep-dive. We went through each position and tenure on my resume and asked for detailed explanations of responsibility on that role
Warning for Other Candidates- The interviews in all three rounds seemed to go very well, and I felt confident I had answered all questions successfully.
However, at the very end of the final round, the VP asked for my previous manager's name. He explicitly stated this was "just for a seniority check" and assured me that they would not contact this person without my consent.
I was rejected shortly after this round. Based on how well the interviews went, I have a strong suspicion that they did contact this reference without my permission, and this may have been the deciding factor.
Advice to Management Please be transparent with your hiring process. If you tell a candidate you will not contact a reference without their explicit consent, please honor that agreement. This creates a more trusting and professional candidate experience.
Sollicitatievragen [1]
Vraag 1
1. how to create lagged features in order to avoid data leakage
2. If we pick single tree from DT, RF, XGB - how will the bias variance look like
3. TF-IDF
4. Weakness , how would manager define you