Figma Machine Learning Engineer interview questions

6 Machine Learning Engineers have shared notes from their Figma interviews, with the questions they were asked in each round. That is 22% of the 27 Figma interview notes on Screna AI.

6 Machine Learning Engineer interview notes

Figma·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Figma ML engineer screen, one coding problem the whole time. Pretty standard graph traversal territory but they wanted you to actually talk through the tradeoffs between approaches rather than just code it up.
  • Given a 2D grid where cells are either land or water, count the number of distinct islands, where an island is any group of land cells connected horizontally or vertically.

“I went with DFS immediately which was fine, but they pushed me on why not BFS or union-find.”

View Post

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

Jun 2026
Figma ML engineer round that ended up being more systems-thinking than I expected. The whole session revolved around one design problem, but it went deep fast.
  • You have a document model (think layers in a design tool) that supports add, delete, move, and property-change operations. Build an undo system that correctly reverts the most recent operation. Walk through how you'd record operations and compare that approach against just snapshotting the whole document.

“I started with snapshots because it felt safe, and the interviewer just kind of waited.”

View Post

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

Jun 2026
Figma ML engineer interview that was more systems-heavy than I expected. The whole session centered on one design problem about undo behavior in a layer editor, and they kept pushing on edge cases I hadn't thought through.
  • Design a batch operation primitive for a layer/document editor where multiple mutations can be grouped and treated as a single atomic unit for undo. How do you structure the operation log, handle nested batches, and deal with partial failures mid-batch?

“Started with the basic command pattern and felt pretty good about it, but then they asked about nested batches and I kind of stalled.”

View Post

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

Apr 2026
ML system design round at Figma for an MLE role. The whole session was one big open-ended question about building a home feed recommender from scratch, and they really wanted you to drive the whole thing end to end without much prompting.
  • Design a home feed recommendation system for a large consumer app. Walk through the full pipeline from problem framing to serving infrastructure.
  • How would you frame the ranking problem as a multi-task learning setup, and what targets would you optimize for?
  • How do you handle freshness and diversity in the re-ranking stage without tanking relevance scores?
  • What does your serving pipeline look like, and how do you meet latency requirements across the retrieval, ranking, and re-ranking stages?

“This is the kind of question where you can either go broad and shallow or try to go deep on one layer and run out of time on everything else.”

View Post

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

Apr 2026
Figma ML engineer interview with a meaty system design question around building a text-to-HTML generation service. Covered a lot of ground, from model training to serving infrastructure, and I felt pretty good about some parts and totally lost on others.
  • Design a service that takes a natural language description of a webpage and generates the corresponding HTML (and optionally CSS and JS). Walk through the full system: model choice and training, alignment, evaluation, and serving.

“This one had a lot of layers.”

View Post

Figma·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Figma ML engineer interview with a coding problem centered on undo/redo behavior in a layer editor. The problem was deceptively tricky once batched operations entered the picture.
  • You're building a layer/document editor (think Figma-like) that supports mutating operations and an Undo command. Implement Redo such that it re-applies the most recently undone operation, handles batched operations as atomic units, and invalidates the redo stack whenever a new mutation occurs after an Undo.

“The basic redo stack wasn't bad, two stacks, push to redo on undo, pop from redo on redo, clear redo on new mutation.”

View Post