Levels and expectations
Mapping titles across companies
Senior at a startup may be mid-level at big tech. Map with scope, autonomy, complexity and influence — not vanity strings.
"Senior" at a 20-person startup may match mid-level scope at a big tech company. "Staff" at one firm may map to senior at another. Comp and leveling sites exist because HR titles are not a shared currency.
Do not negotiate or self-assess on title strings alone. Ask: what is the scope of problems this level owns here?
Concept
The four axes are the translation layer. A title is a sticker. Scope, autonomy, complexity and influence are the job.
Walking through a rough map
People use levels.fyi for orders of magnitude, not gospel:
| Industry shorthand | Amazon-style | Meta | Dropbox framework | |
|---|---|---|---|---|
| Junior | SDE1 | L3 | E3 | IC1 |
| Mid / SDE2 | SDE2 | L4 | E4 | IC2 |
| Senior / SDE3 | SDE3 | L5 | E5 | IC3 |
| Staff | Principal SDE | L6 | E6 | IC4 |
Warning
A "senior" who only closes tickets is mid-level work with a senior sticker. A "mid" who owns ambiguous multi-month delivery is already practicing senior behaviors — which matters for growing scope even before the title moves.
Startups often collapse levels: everyone is "engineer" until they suddenly need a ladder for fairness. That does not mean levels are fake; it means you must describe your work in axis language on your resume and in interviews.
When interviewing, translate your stories into axis language. "I was senior-titled" is weak. "I took an underspecified reliability problem, proposed three options, led a two-quarter migration affecting four teams" is leveled.
What managers actually use levels for
Levels are a shared language for staffing, pay bands and "who should own this?" They are also imperfect. Politics, headcount freezes and weak managers distort them. Still, as an IC you win by making the axes visible in your work:
- Write status as outcomes and risks, not hours.
- Take ownership of the user problem until it is solved.
- Expand scope deliberately: ask for the ambiguous adjacent problem after you deliver the clear one.
- Keep artifacts (design docs, launch notes, metrics before/after) so promotion evidence is not a scramble.
On a healthy team, you should be able to answer: What level am I hired at, what does the next level look like here in writing, and what evidence would convince my manager? If that document does not exist, use a public ladder as a stand-in and align in a 1:1.
Common mistakes
- Negotiating the string. "Make me senior" without a change in problem radius.
- Copying another company's title onto your resume without the work. Interviewers will probe the axes.
- Assuming startup "everyone is an engineer" means levels are fake. Describe the work, not the org chart.
Check yourself
You are "senior" at a 20-person startup and interviewing at a large company. What do you put in the story?
The problems you owned — scope, ambiguity, complexity, who changed because of you — not the title string.
A company maps you to L4. You were "staff" at a six-person shop. What question should you ask?
What problems does L4 own here, and do my recent projects match that radius?
Key takeaways
- Titles differ by company; map with axes and tools like levels.fyi, not vanity strings.
- Senior-sticker plus ticket work is still mid-level work.
- Make axes visible: outcomes, artifacts, adjacent ambiguous problems.
- Align with your manager on the written next-level bar, or use a public ladder as a stand-in.
Further learning
Next, grade yourself with evidence: leveling self-assessment.
levels.fyi — levels and compensation mappingSearch 1M+ data points for different companies, job titles, career levels, and locations. Explore our tools to help you get paid more!Levels.fyi
Related
What the levels mean
Junior, SDE2, SDE3 and staff: scope, autonomy, impact and influence at each level.
The four leveling axes
Scope, autonomy, technical complexity and influence — the four questions every engineering ladder is really asking.
Leveling self-assessment
Pick a public ladder, score three recent pieces of work on the four axes, and mark what you show consistently versus once.