← Back to Directory

Microsoft interview notes, page 11 of 60

FAANG / Big Tech

Most of these Microsoft write-ups come from Software Engineers (372), then Product Managers (124) and Machine Learning Engineers (53). The most reported rounds are the technical phone screen (255), the product sense round (111), the system design round (93) and the behavioral / leadership round (49).

597 interview notes · updated Oct 2026

Microsoft·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Microsoft ML Engineer interview round focused on LLM prompting and post-training, mixing live coding with the OpenAI API and conceptual discussion about fine-tuning methods. Pretty dense for a single session.
  • Write code that calls the OpenAI Chat Completions or Responses API to perform a task, including proper prompt structure with system and user messages.
  • How would you design a few-shot prompting setup, and what makes a good few-shot example selection strategy?
  • Compare supervised fine-tuning, RLHF, and DPO as post-training techniques. How does each one shape model behavior differently?
  • What does instruction tuning do to a base model, and how is it different from general supervised fine-tuning?
  • How would you evaluate a fine-tuned language model, and what safety considerations would you factor into the process?

“I knew the API reasonably well but fumbled the exact parameter names under pressure.”

View Post

Microsoft·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Microsoft ML Engineer interview, technical phone screen focused entirely on LLM fundamentals and deployment. Two big questions, no fluff, they went pretty deep on both. Felt like a solid screen for someone who's actually shipped models versus someone who just reads papers.
  • Walk me through the Transformer architecture: what problem it solves compared to RNNs, and explain the core components including embeddings, positional encoding, self-attention with Q/K/V, multi-head attention, feed-forward layers, residual connections, and layer norm. Also, what's the computational complexity of full self-attention with respect to sequence length?
  • You're deploying an LLM-powered feature like an internal assistant or customer support bot. What are the main real-world challenges, what architecture would you propose to address them, and how would you evaluate it before launch and monitor it afterward?

“This is the kind of question where you think you know it until you're saying it out loud.”

View Post

Microsoft·Machine Learning Engineer·Onsite - System Design / Architecture

Apr 2026
A deep system design round at Microsoft for an ML engineer role, focused entirely on building a video VLM from the ground up. The scope was massive and they expected you to connect data, modeling, training, and eval into one coherent story, then tie it back to something you'd actually shipped.
  • Design a video Vision-Language Model end-to-end, covering data sourcing and preprocessing, model architecture, training strategy, and evaluation. Then walk through a past project and explain what you'd do differently with this design.
  • How would you source and build a training dataset for a video VLM, including captioning pipelines, multimodal alignment data, and filtering or deduplication?
  • What are the trade-offs between different video encoder architectures like ViT with frame sampling, 3D CNNs, and perceiver-based models for a video understanding task?
  • Walk through the pretraining objectives you'd use for a video VLM, such as contrastive learning, captioning, or masked language modeling, and how you'd stage the training.
  • How would you evaluate a video VLM across retrieval, captioning, and visual question answering benchmarks, and where does human evaluation fit in?

“This was basically four questions in a trench coat.”

View Post

Microsoft·Product Manager·Hiring Manager Screen

Apr 2026
Interviewed for a PM role at Microsoft, just one question from what was shared here. Pretty introspective stuff, not the product metrics grilling I expected.
  • What motivates you to get out of bed every morning?

“Wasn't ready for this to be the opener.”

View Post

Microsoft·Software Engineer·Technical Phone Screen

Apr 2026
Microsoft SWE interview with a graph/union-find problem that sounds straightforward until you're actually coding it under pressure.
  • You have an m x n grid that starts as all water. Given a sequence of operations that each flip one cell to land, return the island count after every single operation. Islands are 4-directionally connected. Solve it efficiently.

“I knew union-find was the right call pretty quickly, which felt good.”

View Post

Microsoft·Product Manager·Onsite - Product Sense / Strategy

Apr 2026
PM interview at Microsoft with a couple of product sense questions focused on their own tools. Pretty conversational, nothing too intense, but the questions were deceptively open-ended.
  • Have you used PowerPoint before, and what would you change about it?
  • What recent changes or trends have you noticed in PowerPoint?

“Classic Microsoft 'critique our own product' question.”

View Post

Microsoft·Software Engineer·Technical Phone Screen

Apr 2026
Microsoft SWE interview with a graph/topological sort problem. Pretty standard stuff for this kind of role but the cycle detection piece is where things get interesting.
  • Given a number of courses and a list of prerequisite pairs, return a valid order to complete all courses. If no valid ordering exists due to a cycle, return an empty array.

“Topological sort, which I knew going in, but I fumbled the cycle detection part longer than I should have.”

View Post

Microsoft·Technical Program Manager·Recruiter / HR Screen

Apr 2026
Interviewed for a TPM role at Microsoft, just one question about what gets you excited about tech. Short and vague, not much to say.
  • What excites you most about technology?

“Pretty open-ended opener.”

View Post

Microsoft·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Microsoft ML Engineer interview with a two-part question covering LLM decoding theory and a live Python implementation. The coding portion was more involved than I expected for what seemed like a conceptual topic.
  • Compare greedy decoding, top-k sampling, top-p (nucleus) sampling, and beam search for LLMs. What are the trade-offs in terms of determinism, output diversity, fluency, and handling low-probability tokens? When would you choose one over another?
  • Implement greedy decoding in Python. You're given a function that takes the last token and returns a probability distribution over next tokens as a list of (token, probability) pairs. Generate a sequence by always picking the highest-probability next token, stopping at an end-of-sentence marker or max length, and return both the generated sentence and the product of the chosen probabilities.

“I felt okay here but rambled a bit on beam search.”

View Post

Microsoft·Software Engineer·Onsite - Multi Round

Apr 2026
Microsoft SWE onsite with two parts back to back: behavioral first, then a coding problem. Nothing too wild but the combo of both in one sitting kept me on my toes.
  • Walk us through your background and a past project you're proud of.
  • Given a grid where cells are either walls (-1), gates (0), or empty rooms (a large max integer), fill each empty room with the distance to its nearest gate. If a room can't reach any gate, leave it as-is.

“Pretty standard opener.”

View Post