Amazon

After MLSS ' 26 got over then we were given a hiring interest form. Round 1 - OA Round 2 (interview round 1 called as sde round): In my case it was taken by sde -1 at amazon. It consisted of :- 2 DSA Problems + no LPs Merge Intervals Djikstra with path sum Interviewer was really supportive lasted for around 55 mins I was asked to completely code the first question and just explain the approach for the second along with proving the time complexities for both questions. Note -LPs can be asked as they were asked by some people in our cohort(MLSS'26). Round 3: taken by an applied scientist at amazon this round was called ml round (depth+breadth). ML Questions only Questions related to resume project. Interviewer was very supportive and I was grilled on resume project explaining its architecture , why I used a particular loss function, CNN Vs Transformer , some transformers related problems , then methods to reduce overfitting , decision trees : xgboost and light-gbm comparision(detailed). He asked regarding project till 1hr then he moved on with ml questions. But, he was asking too many cross-questions at one go like 3 or 4 questions at a time which you can only answer if u know ur project really well. Tips: For R1 : Prepare DSA very well. There is no language restriction, you can code in any language of your choice. Questions asked to peers were also same like 1 graph(rotten oranges) + 1 simple or 1 DP (House robbers)+ 1 simple or 1 binary search(minimise the maximum) + 1 simple . So, dont skip any topic. For R2: Be ready to answer everything regarding the project . You should know the architecture , no of params (I was asked to calculate no of parameters in live interview and prove), you need to prove why you took a particular decision of choosing this over that related questions which will then lead to some more detailed ML Questions (that's the general flow).