Interview Questions

Anonymous Halibut - "The product lens I view the mousetrap through is a product which has a decision maker (someone who wants to catch a mouse) and an end user (yes, yes, the mouse to be trapped). I believe that the mousetrap is superior when it shows a better speed of catching mice. So given that there are mice to catch, it has a better 'conversion' rate of mice which took interest in the mousetrap / mice who got caught - this is a decision maker, e.g. a homeowner, perspective For a mouse, a trap is better when it"See full answer

Rahul J. - "Target market and pain points to be solved Impact of solving the problem: TAM Existing solutions and their problems Value proposition of my product Costs and resources required in building and maintaining"See full answer

Exponent - "We've identified the problem as a Design a Product question. Use the following framework for tackling these types of questions: Ask Clarifying Questions Identify users, behaviors, and pain points State product goal Identify current solutions Brainstorm new solutions Evaluate solutions Measure success Summarize We'll go through each of these step by step. Ask Clarifying Questions The PM interview isn't about your ability to come up w"See full answer

Danyal R. - "Understand the problem: Chipotle is a chain, located across the US. It's in every major city Make Assumptions I'm assuming in each state there's approx 5 Chipotle restaurants. So in total across US there is 5 * 50 = 250 Chipotle's Let's calculate the sales revenue for a whole year. Create an equation Total Yearly Revenue = # of people going to Chipotle X* *frequency X cost of a meal Break down the problem \# of people going to Chipotle There's about 300M in the US. Out of 300M I'll a"See full answer

Sourav K. - "We can use a pool of memory where each index we can store the parent of current index then we can use same lca approach for array instead of pointers."See full answer

Anonymous Hornet - "label #shape = [batchsize, posids] X #shape = [batch_size, ch] hidden = nn.Linear(batchsize, ch) #shape = [batchsize, ch] logits = nn.Linear(ch, C) #shape = [ch, C] #modifications start here neg_ids = torch.randint(0, C, 0.1 * C) #e.g. if we want to randomly sample only 10% of C outputneg = np.dot(logits[:, negids], hidden(X)) #shape [batch_size, C1] outputpos = np.dot(logits[:, posids], hidden(X)) #shape [batch_size, C2] output = torch.concat(outputpos, outputneg) #concatenate out"See full answer