Insights from our first free mock interview (8/7/7; Engineer)

Insights from our first free mock interview (8/7/7; Engineer) We posted a couple of days back how we would take free mock PIs as a way of giving back to the community. As our bandwidth to take interviews would be limited, we will also share insights from the interviews we take, so they are generally beneficial to the whole community. Profile of the guy we interviewed: 8/7/7; Engineer (IT) Native to Nagpur; 5 months work-ex as a data engineer **Specific pointers**: **1. Introduction** **Things not to do** * Do not make your intro long-winded, chronological or fact based, meaning you should not start off by listing where you were born, the fact that you were a topper all through your school life, then you picked up engineering, interned somewhere and then have started working somewhere now. * This does not work because it is too long (panelists will cut you off after 30-45 seconds), it is a collection of facts (what do I do with the fact that you were a topper) and is chronological (your achievements in school matter much lesser than what you are upto right now). Further, it does not seed anything in the interviewer's mind - where do you want me to take this interview? You did not emphasise anything in particular. **Things to do** * Keep your intro restricted to 30-45 seconds, and make it narrative driven - try to emphasise things you want the interviewer to pick up. * For instance, in this case, the intro could have been around how the guy was born and brought up in Nagpur, where he saw small business struggling because they would take calls based on intuition - this is why he got into data - and now that he has a strong fundamental knowledge, he wants to move over to the business side of things (this is my take based on my conversation with the candidate). * Now you would obviously expect the panel to test you on technicals related to data/stats/modelling or how exactly data translates to business, ask you about Nagpur, or go on a different tangent altogether (interviews can be random). But, the first 2 things are what you should be prepped for. **2. Why MBA** **Things not to do** * Do not give a first-level answer and expect no cross-questioning. For instance, if you want to transition from technical to the business side of things, what exactly will that entail? Have you seen something like that happening at work? Have you taken the initiative to pick up some of those projects? **Things to do** * Be super specific in your answers + back them up with examples. Generic statements do not work. For instance, is there a project which you were a part of, where you got to experience the business side of things? What was the overall context of the project, what impact did you drive, how did you contribute? **3. Static GK** **Things not to do** * Do not walk in without having a basic context on things that you should probably know. For instance, if you are from Nagpur, you should know that it is famous for oranges, and was the place where the zero milestone was placed (follow-ups could be what does it signify, which archaeological survey was it a part of, or is it still relevant now). **Things to do** * Be prepared for static GK - we have written a post on things you should cover and how to structure your prep for static GK and general awareness - you can look it up. **4. Do not skip academics, especially if it relates to your work** **Things not to do** * Do not skip your core UG subjects, or things which relate to your work and you should know. For instance, data modelling, having context on running an ETL pipeline, figuring out what would be predictor and dependent variables for the model, multicollinearity, categorical variables and what their coefficient represent, are probably things you should walk in expecting you would be grilled on, if you are working as a data engineer. **Things to do** * Go beyond academics and try to link concepts to business applications. For instance, in this interview, we asked how the candidate would use data from MagicBricks to build a price prediction model for real estate. What data would you extract? What would you predict? How would you build the pipeline? How would you solve the issue of multi-collinearity among independent variables? What insights would you give to the builder? Typically, you would use a hedonic price regression model for this, where you would predict the price/sqft based on variables such as the locality the flat is in, gated/un-gated society, amenities like swimming pool (yes/no - categorical variable), how far off it is from schools/hospitals, whether it is unfurnished/furnished, etc. * Insights could be price premium for premium locality is justified, how much incremental value adding a pool to the locality means, whether buyers prefer furnished or unfurnished flats and does this differ based on the type of locality (premium/non-premium). All the best! Hope this was helpful. Will post more transcripts from the mock interviews we do.