Microsoft Fresher Interview Questions

“I would treat each cell as a node in a directed graph, with an edge from a cell to every cell whose formula uses it. When a value changes, only the cells reachable from it need work, and they must be done in topological order, so a cell is computed after everything it depends on. A depth-first search that records when each node is finished gives that order. If the search reaches a node that is still on the current path, there is a cycle, which is a circular reference. In that case I would stop, mark the cells involved and show the user a clear message rather than loop forever.”

“I would keep two stacks. Every edit is stored as a small operation, such as insert this text at this position, with enough detail to reverse it. Doing an edit pushes it onto the undo stack. Undo pops it, reverses it and pushes it onto the redo stack. Redo does the opposite. When the user makes a fresh edit after undoing, the redo stack is cleared, because those futures no longer apply. Storing every single key press would make undo tedious, so I would group letters typed quickly into one operation. I would also cap the history so memory does not grow without limit.”

“The worst outcome is quietly losing someone's work, so I would not let the last upload win. Each device records the version it started from. When both upload with the same starting version, the server knows there is a conflict. For plain text a merge may be possible, but for most files the safe choice is to keep both: save the second one with a name that says it is a conflicted copy from that device, and notify the user on both devices. They can then compare and choose. If only one device changed the file, there is no conflict and the newer version simply replaces the older one.”

“Each program is given its own virtual address space, so it thinks it has the whole memory to itself. The operating system splits memory into pages and keeps a table mapping each program's virtual pages to real physical pages. When physical memory runs out, the system moves pages that have not been used recently to a file on disk. If a program touches such a page, a page fault happens, and the system brings it back, possibly moving another page out. Disk is far slower than memory, so when many programs keep needing pages that are on disk, the machine spends most of its time swapping, and everything crawls.”

“In your project, a timetable tool used by my college department, I changed the export from a plain text file to a formatted spreadsheet because it looked nicer. A week later a staff member told me her script that read the old file had stopped working. I had never thought anyone processed the file. I apologised, brought the old export back as a second option, and kept the new one as the default. Then I asked the other users how they used exports before touching it again. Since then I list who uses a feature before I change it, and I keep old behaviour available for a while.”

“During an internship on your project, I was asked to host our web app on a cloud service, which I had never used. I gave myself two days to learn only what the task needed: hosting the app, a database and a way to deploy. I followed the official beginner guides, built a tiny test app first, and kept a notes file of every command. When I got stuck on permissions, I asked a senior engineer a precise question instead of a vague one. The app was live within the week and your result. I learned to learn the smallest useful part first.”

“I would start by believing the tester and asking for exact details: the system version, memory, the error text and the steps. Then I would try to reproduce on a similar setup, using a virtual machine with the older version if I do not have the laptop. Common causes are a library that behaves differently on the older version, a feature that does not exist there, or my fix using more memory than a small machine has. Once I find it, I fix it in a way that works on both, test on both and add the older setup to my checklist. Works on my machine is not finished.”

“I am interested in software that has to keep working for people for years, not just impress them once. I have used Word and Excel since school, and I have seen features change while my old files still open. I want to learn how that is done. In your project, I had to keep old data files working after I changed the format, and I found the problem harder and more interesting than adding new features. I would like to work on tools that people rely on for their work, and learn from engineers who have kept software dependable through many versions.”