How Hiring Decisions Get Made

Notion has been open about the two questions it is really asking in a loop. The first is whether you excel at your craft. The second is whether you share the company's values, where "kind and direct" is the pairing people quote most from its published set. Almost everything in the interview is a way of getting at one of those two things, and the tone is set to match. Notion frames its interviews as two-way conversations rather than interrogations, and it weights reference checks unusually heavily. That combination tells you how a decision comes together here: less a scorecard tallied in isolation, more a small group deciding whether they want to build alongside you.

The coding and design rounds are shaped by what Notion is. Everything in the product is a block: a typed unit with an id, a type, a properties object, an ordered array of child blocks, and a parent pointer. Those blocks form a recursive tree that syncs live across collaborators and keeps working offline. So the coding round tends to hand you a small feature over a document, a table, or a block tree, and the design round is named High-Level Design and Data Modeling for a reason. It wants a clean schema and efficient queries over something Notion-shaped, not a capacity-planning exercise. Reasoning in blocks, permissions, and sync is the whole game, and defaulting to a design-canvas or scene-graph mental model is a miss, because Notion is documents and databases, not a pixel surface.

The rest of the machinery is kept quiet. Notion does not publish a rating scale, an internal level ladder, or a named committee structure, and a fair amount of confident detail online is either inference or stale. One fact worth pinning down before anything else: Notion is a private company. There is no ticker and no public filing. The most recent mark is roughly eleven billion dollars from a late-2025 employee tender, and any page claiming a hundred-billion valuation or a filed S-1 is fabricated. That status matters for how you read a level, because at a private company the equity is the whole comp story and it is illiquid paper, not a cash number. This lesson separates the parts of the process that hold steady from the parts Notion keeps private, and it does not invent precision Notion has never offered.

The Shape of a Verdict

There is no official scale to quote, so read the four bands below as a rough feel for how a round tends to come back, not a rubric Notion stands behind. What each round gets measured against is the sharper thing. In coding, the base task is usually easy and the follow-up is the real interview. Working and tested beats clever and broken, and stating an invariant before you are asked, like a fresh insert or delete clearing the redo history on a text editor, reads well. In the data-modeling round, did you land on the block schema and write queries that walk the tree efficiently. In system design, did you split the server-authoritative online path from the offline local-first path and pick a defensible conflict story. And threaded through all of it sits the read on craft and product judgment, because that is the first of the two hiring questions.

The Notion angle

The base problem is easy on purpose, and the follow-up is where the round is decided. Notion's coding loop leans practical and product-shaped, so you will often start somewhere simple, count and sum per group over a table, or complete a few stubbed functions on a small text editor. Passing the visible tests is necessary and nowhere near sufficient. The pass is earned on the escalation. Make the aggregation incremental so it updates on write instead of recomputing. Clear the redo stack when a new edit lands. Reconstruct the block tree from flat parent ids and handle the orphans. Some teams still ask a plain medium or hard algorithm question, so do not assume every round is a feature build. When it is, budget your time for the follow-up rather than sprinting to a first green result.

What a Strong Round Looks Like

There is no published pass rule, so the contrast below describes how loops tend to land. The through-line is that clean execution is the floor, and product judgment plus a real feel for the block-and-sync domain are what separate a clear yes from a miss. Read these as the gap between a round that comes back an argument for you and one that comes back a concern.

The Notion angle

A design round that stays in document-and-block terms is the credibility flex, and a canvas reflex is the common miss. Notion is deliberately not a design surface. It is a document made of nested typed blocks, with databases whose rows are blocks and permissions that inherit down the tree. Everything is server-authoritative when you are online and merged through a CRDT when you are offline. When a candidate defaults to a scene-graph or cursor-engine framing, it signals they are answering a different product's question. Anchor every design in the block schema, name the parent pointer as the permission spine, and treat offline local-first sync as a first-class path rather than an afterthought. That is the reasoning the round is built to reward.

Who Holds the Decision

The honest description is small and hiring-manager-led. Notion is around a thousand people, so there is no evidence of a Google-style independent cross-org committee reading written packets in another building, and you should not prepare for one. A hiring manager is typically engaged early, often for a conversation before or around the technical screen. After the loop, the interviewers write up their feedback, and the hiring manager and recruiter consolidate it against the two hiring questions and the values. Reference checks carry real weight in that consolidation, more than at most companies. The people who can speak to your work are part of the decision, even though they never sit in a round. Notion does not disclose the exact debrief structure, so treat the mechanism as genuinely opaque and prepare for the part you control: leave every interviewer something concrete to write down.

The Notion angle

Craft is graded as an ethos, not counted as one of the values, and it runs through the whole loop. The company has been explicit that hiring turns on whether you excel at your craft and whether you share its values. Craft here means the quality of what the user feels, not only clean code. That shows up as product taste in the behavioral round and as care in how you model a block schema or handle an edge case in coding. Do not miscount craft as a fifth named value, and do not paraphrase the published values into generic slogans. The safe, durable thread to build around is kind and direct: show it by taking feedback in the room without defensiveness and by disagreeing plainly and warmly when you do.

The Level Ladder, and What Calibrates It

The process lesson walks the loop; here is the ladder the decision maps you onto. Notion does not publish its internal engineering titles. So read the L1-through-L5 framing below as the normalization public compensation data uses, not Notion's own names, and treat the scope and experience minimums as the real content. The senior band is the center of gravity, and the top rungs are the least precise. Your level is calibrated through the loop and read off the scope you demonstrate, especially in the data-modeling and design rounds, not off the title on your resume. No comp is attached to it here. That last point matters more than at a public company, because the level sets an equity band that is illiquid private stock, and the offer lesson is where that gets priced.

The Notion angle

If you are near a level boundary, point your prep at data modeling and system design over more coding reps. Clean coding is a gate you have to clear, but it rarely moves you up a rung. Senior scope gets read in the block-schema conversation and the collaborative-and-offline design round. A candidate who codes cleanly and then stays generic on modeling is the classic profile that lands a level low. The stakes are higher than usual because leveling drives a large equity delta, and at a private company that equity is the majority of a senior package. So once your feature builds run reliably, shift the marginal hour to articulating the block tree, permission inheritance, and the online-versus-offline sync split, and to a project you can describe with sharp, quantified ownership.

After the Loop

The full process usually runs four to six touchpoints over roughly two to five weeks, with an average time-to-hire around three weeks; Notion publishes no official window. Because the company is small and the hiring manager is involved early, team fit is assessed as part of the loop rather than assigned by a separate central match after you clear a bar. The last stage often includes a team and references step, which is where the heavy weighting on reference checks lands. Read a quiet week as scheduling rather than a signal, and keep your recruiter warm through it, since a small company has fewer moving parts but also fewer people to chase a stalled thread.

• Recruiter screen to decision: ~2 to 5 weeks. Typical end to end across four to six touchpoints, with a reported time-to-hire near three weeks. Variable by team and scheduling.
• Team fit: Part of the loop. A small, hiring-manager-led org assesses team fit and references through the loop rather than matching you to a team afterward via a central committee.
• Decision process: Debrief consolidation. Interviewers write up feedback that the hiring manager and recruiter consolidate against the two hiring questions and the values. The exact structure is not public, so leave every interviewer something concrete to write down.

If the answer is no, ask which themes to work on. Recruiters rarely hand over round-by-round notes, but they will often pass along what the loop flagged. This decision turns heavily on the follow-up in coding and the block-model reasoning in design, so that steer is the most useful thing you leave with. A no is not always permanent either, since a fast-growing company's teams and needs shift, and Notion documents no reapply cooldown, so anyone quoting a fixed window is guessing. Ask your recruiter to confirm the timing for your case. The next lesson turns all of this into a concrete plan.

Common questions