OpenAI Interview Questions

Robert H. - "The addition of an intermediate "sanitization" ML is something Neeraj used in the Uber Eats design and again seems kind of outside the scope here and redundant. This can simply be built into the AI response model to save a step. It's not clear what benefit this step provides, and he basically said we should have it "just because it would be good" and there's no concrete reasoning why to include it. Adding a Kafka queue to handle the thumbs-down ratings? For what purpose do you need a queue othe"See full answer

Shraddha D. - "I like chatgpt for the following users Getting industry references are easy and time saving Getting recommendations is very easy Responses are accurate I use chatgpt to get feedback on content and identify gaps in my documentation or thought process I use chatgpt as a search engine and to get conscise situation based information. Chatgpt offers varierty of other tools in the explore version where similar users can create different content PRD template Chatgpt users 1."See full answer

Srikhar S. - "t"See full answer

Dessalew A. - "Gradient Descent = core engine for training most ML models It works by iteratively minimizing loss via gradients Many improvements exist (Adam, RMSProp, etc.) Alternatives exist for: Faster convergence Non-differentiable problems Direct metric optimization"See full answer

Sachin P. - "The key here is a) not to overreact to Sales bypassing me going straight to Engineering (while understanding every team acts according to their incentives), and b) communicate with all parties as much as possible to get everyone on the same page. 1) First, I'd speak to Sales and understand the feature being requested. I consider Sales to be one of the most important primary research funnels - they have direct, regular touchpoints with the client and are finely attuned to the lifeblood o"See full answer

Snehangsu B. - "In a recent implementation of a large core banking project where we were breaking away from monolithic architecture to microservice based one we faced a situation where customer wanted to have a single repository of user and role management. Since Micro service based architecture talks about individual entities for multiple product suite it was mandatory to maintain individual user base. Based on the customer feedback, we had to come up with an unique design of single user base and replicating"See full answer

Scott S. - " The Situation A few months ago, our trading platform started experiencing significant latency issues during peak trading hours. This latency was affecting our ability to process real-time market data and execute trades efficiently, potentially leading to substantial financial losses and missed opportunities. Identifying the Problem The first step was to identify the root cause of the latency. I organized a team meeting with our data engineers, DevOps, and network specialists to gather"See full answer

Krish M. - "Situation: I led a critical initiative to re-architect the high volume customer data platform. The system handled over 90 PB of records handling around 70 billion customer records daily and was foundational to customer data capabilities across multiple enterprise products. The existing architecture faced significant challenges around latency, accuracy, and scalability, and any disruption would directly impact global customers and downstream systems. Task: I was responsible for leading t"See full answer

Anonymous Capybara - "using a relational database isn't a good choice for this system! we need more availability here than consistency (CAP theorem)"See full answer

Darpan D. - "Clarifying questions and Assumptions ChatGPT search means the search function inside the chat app? OR ChatGPT search Chrome extension? Assumption: Search inside the chat app. Is there any location restriction in this analysis? Assumption: USA only. Is there any user segment restriction in this analysis? Assumption: All user segments. Are we assuming the ChatGPT search already exists or going back in time before the ChatGPT search existed? Assumption: Go back in time"See full answer

Aditya S. - "I told a story about how our team was focussed on moving a key metric i.e. NPS and to do that we build 3 top requested user feature. Post release the detractors % didn't move even though the detractors request for features shipped went down. Then I connect with users and did some analysis post which we realised that we need to pivot our focus from shipping features to enabling complete workflows for our users i.e. shipping all those feature which are used together in a feature as then only users"See full answer