Uber's performance review process for software engineers
Uber review coming up?
Get a plan to walk in with evidence, not guesses.
At most big tech companies, L4 is a comfortable resting spot. Mid-level engineers can spend years there, doing solid work, drawing solid comp, without anyone expecting them to sprint toward senior. Google L4, Amazon SDE II, Meta E4: these levels have long been understood as sustainable long-term positions for engineers who don't want the staff track.
Uber is different. L4 (Software Engineer II) is not a terminal level. The expectation: engineers who aren't progressing toward L5 (Senior Software Engineer) within three to five years can find themselves managed out. That changes how you should think about performance reviews at Uber from day one.
This guide covers how Uber's review process actually works, what happens in calibration, what the promotion committee looks for, and what separates engineers who advance from the ones who stall.
Uber's engineering level structure
In March 2022, Uber renamed its engineering levels to match industry conventions. The old L5A and L5B split, which had grown confusing internally and externally, was retired. L5B became Staff Engineer (now L6), and a new Level 9 was added at the top.
| Level | Title | Notes |
|---|---|---|
| L3 | Software Engineer I | New grad / entry-level |
| L4 | Software Engineer II | Mid-level; not a terminal level |
| L5 | Senior Software Engineer | Previously called L5A |
| L6 | Staff Engineer | Previously called L5B; cross-team scope required |
| L7 | Senior Staff Engineer | Org-wide scope |
| L8 | Principal Engineer | Strategic, org-defining impact |
| L9 | Distinguished Engineer | Extremely rare; added in 2022 |
If you joined before March 2022, you'll encounter references to L5A and L5B throughout older Team Blind threads. For all practical purposes: L5 is Senior, L6 is Staff. The level numbers stayed the same; only the naming conventions changed.
How Uber's performance review cycle works
When reviews happen
Uber runs reviews twice a year, tied to the first half (H1) and second half (H2) of the calendar year. The H1 cycle closes around June; H2 closes around December. Bonuses are calculated per half based on performance outcomes from that cycle. Engineers who are new to Uber are not eligible for promotion until they have been at the company for at least one year.
The review flow
Each cycle follows roughly the same sequence:
- Self-assessment: you write a written evaluation of your contributions and scope for the cycle
- Peer reviews: you nominate three to five peers to provide written feedback
- Manager evaluation: your manager writes an assessment and, if applicable, submits a promotion nomination
- Calibration: managers in your sub-organization meet to align on ratings and nominations
- Executive review: the Engineering executive reviews tentative decisions for the full organization
- People team alignment: final decisions are confirmed with HR before being communicated
Uber's review system is prose-based. You will not see a numerical score at the end of the process. What you will see is where you land relative to peers in calibration, reflected in your compensation outcome. Internally, this maps to an implied performance tier, but the specific rating label names Uber uses are not publicly documented. Amazon takes a different approach: a named five-tier system with fixed percentages, detailed in the Amazon performance review guide.
Calibration: where the real decisions happen
Calibration at Uber runs in multiple rounds, each adding a layer of scrutiny. For a deeper look at how calibration works across big tech, see how performance review calibration works.
First, managers in a sub-organization (typically five to eight managers plus a director) meet to discuss each person's tentative rating and any promotion nominations. Then the Engineering executive reviews decisions for the full organization, with focus on promotions and top and bottom performers. Finally, the executive aligns with the People team before outcomes are finalized.
From Gergely Orosz, who ran calibration as an Engineering Director at Uber:
"In an organization with several hundred engineers, a common calibration process looks like: managers submit tentative ratings and promotion decisions, then managers in a sub-organization meet together in a group of 5-8, including the director, to discuss each of the tentative decisions."
Your work doesn't walk into calibration by itself. Your manager presents it. Their ability to articulate your impact, and their credibility in that room, directly affects how your case lands. As one Uber engineer put it on Team Blind, promotions "heavily depend on your Engineering Manager, your skip-level manager, and the budget allocated for promotions."
Promotion committee (for L5 and above)
For promotions to Senior (L5) and above, Uber runs a separate promotion committee review on top of the standard calibration process. Senior engineers and managers on the committee review the full promotion packet and make the final call.
This adds another round of scrutiny and another set of stakeholders whose bar you need to clear. Engineers who've been through the process describe the committee as unpredictable: you can go all the way to the committee having prepared a strong packet, received positive peer feedback, and still get rejected. Cases that look solid at the team level don't always survive committee review.
What goes into a promotion packet
To be nominated for promotion, you need to assemble a packet. Your manager helps guide the process, but you are responsible for the content.
A typical promotion packet at Uber includes:
- Your self-evaluation, written in prose, covering accomplishments and scope
- Peer feedback from three to five peers, also written in prose
- Your manager's evaluation and formal nomination
- Impact evidence: links to design documents (high-level designs, low-level designs), pull requests, shipped features, and quantified business outcomes where possible
- A running impact document covering the cycle, or multiple cycles if building toward staff
The peer feedback component gets underestimated consistently. Engineers tend to nominate peers who know them well, which usually means people on their immediate team. At senior levels, the committee wants evidence that your work had reach beyond that team. Peer reviewers who can speak to cross-team impact carry more weight than five internal endorsements saying roughly the same thing. The guide on how to write peer feedback that holds up in calibration describes exactly the kind of specific, behavioral evidence the promotion committee is looking for.
How to build a case that survives calibration
Document throughout the year, not at review time
Uber's promotion committee sees packets from engineers across the organization. If you're not sure how to structure your self-assessment, see our software engineer self-review guide for a complete framework. Vague impact narratives lose to specific ones. Engineers who build cases that land have been running a live impact document throughout the year, updating it as work ships.
What to track:
- Projects you drove from start to finish, with concrete outcomes (latency improvements, cost savings in dollars, reliability gains, number of users affected)
- Technical decisions and tradeoffs you documented that other teams referenced or adopted
- Cross-team work where your contribution was visible to the other team's leadership
- Mentoring contributions, code review patterns, and knowledge-sharing that helped other engineers ship faster
Ruben Geerlings, a software engineer at Uber who wrote about building impact there, was clear about the framing: describe contributions in terms of customer and business outcomes (gross bookings, cost savings in dollars), not just what you built technically. The committee cannot infer business impact from a technical description.
Work outside your team's charter
Uber rewards engineers who identify and solve problems that go beyond their assigned scope. Management is generally willing to support time spent on engineering problems that benefit Uber more broadly, even when those problems aren't on the official roadmap.
This is also how cross-team visibility develops organically. Working on a problem that touches another team's systems gets you known in that organization without requiring a formal cross-functional project.
For L4 to L5 promotions, team-level complexity is the bar. For L5 to L6 (Staff), cross-team scope is required. Engineers who don't expand their reach beyond a single team tend to stall at the senior level.
Keep writing code
This point comes directly from Joakim Recht, a Distinguished Engineer at Uber, and it applies at senior and staff levels: engineers who stop writing code lose credibility. Designs become less connected to implementation reality. Influence in technical decisions erodes.
As Recht put it: "If you're not writing code, you lose touch with the system...designs will be more decoupled from reality."
The engineers who advance to staff and above at Uber are writing code alongside directing others, not instead of it.
Have the promotion conversation early
Engineers who feel blindsided by review outcomes are usually the ones who never had an explicit conversation with their manager about promotion timing. Before review season, that conversation needs to establish:
- Whether you are being considered this cycle or the next
- What specific evidence would strengthen your case in calibration
- Whether there are budget or quota constraints affecting your org this cycle
Budget is the variable that surprises people most. Uber caps the number of promotions per cycle in each organization. A strong case in the wrong quarter, or in an org where several other engineers are also in consideration, can result in waiting another cycle. Knowing this ahead of time lets you plan. Stripe's performance review process works the same way.
Common mistakes at Uber
Optimizing for activity over impact. Uber has an engineering metrics dashboard that tracks diffs submitted, code review participation, and focus time. These numbers are visible to your manager and director. The temptation to optimize for them is real. But the promotion bar is about business impact, not raw output. Engineers who close a lot of tickets but cannot articulate how their work moved a business outcome consistently underperform in review.
Not maintaining an impact document. The review process is prose-based. If you haven't been tracking your accomplishments throughout the half, you're writing your case from memory. Your manager is in the same position. The engineer whose manager walks into calibration with a specific, well-documented impact narrative has a structural advantage over the one whose manager is reconstructing the year from Jira tickets.
These two mistakes are the most common. The next two are process mistakes, and they're more fixable because they're about timing, not effort.
Choosing peer reviewers only from your team. An all-within-team peer review set reads as limited scope to the promotion committee, even if the individual feedback is strong. In the months before review season, build working relationships and visible contributions with people outside your immediate team.
Waiting to have the promotion conversation. If your manager doesn't know you're targeting a specific cycle, they may not be positioning your case during the months before calibration. Manager advocacy in calibration builds up over time. It doesn't happen in the week before ratings are due.
How long does it take to get promoted at Uber?
Most engineers spend 1.5 to 2.5 years at L3 before reaching L4, then another two to three years before Senior (L5). Getting to Staff (L6) takes three to five or more years after Senior, and is rare enough that it shouldn't be your planning baseline. The full trajectory from new grad to Senior runs four to six years for most engineers.
| Transition | Typical Timeline | Notes |
|---|---|---|
| L3 to L4 | 1.5 to 2.5 years | New grads typically target around two years |
| L4 to L5 (Senior) | 2 to 3 years | L4 has an implicit clock; 3-5 years without progress is a flag |
| L5 to L6 (Staff) | 3 to 5+ years | Very selective; historically rare even in large orgs |
| L3 to L5 total | 4 to 6 years | Around three years for top performers |
Uber has historically promoted faster than some large tech companies. The bar tightened post-2022. In the 2023 review cycle, Uber used its performance review process to identify bottom performers and separate them from the company, treating it as an alternative to formal headcount reductions. If you receive a rating below expectations at Uber, what to do after a bad performance review covers how to respond and rebuild before the next cycle. The review process at Uber carries consequences that go beyond promotion decisions.
The staff level bar (L6) is worth understanding specifically. According to reporting from the Pragmatic Engineer, in a 300-engineer organization, only one person was promoted to Staff over a four-year period. That is not a bar you clear by doing your job well at the senior level. It requires consistent cross-organizational scope and documented impact that others in the organization can point to.
Building your case between now and the next review window
If your next cycle is three to six months out, here is where to focus.
Open your impact document today if you haven't already. Write down every project you've contributed to this cycle, the outcomes, and who outside your team was affected. Be specific: latency numbers, cost savings, reliability improvements, the number of engineers who adopted a library you built.
Identify two or three engineers outside your team who could speak to your work in writing. Not to ask for a favor now, but to make sure those working relationships are solid enough that they could describe your contributions specifically.
Have the explicit promotion timing conversation with your manager. This cycle or the next? What would strengthen the case? Is there a budget constraint to be aware of?
Then document the next three months with calibration in mind. The technical decisions, cross-team contributions, and business impact evidence you generate between now and the close of the cycle are what your manager presents. Make it easy for them to walk in prepared.
Build your case with CareerClimb
Uber's review process rewards engineers who document their impact throughout the year, not the ones who reconstruct it in the week before the self-assessment deadline.
CareerClimb is an AI coaching app that helps you log wins as they happen, map your contributions to your company's promotion criteria, and walk into every review cycle with your evidence already organized. Your AI coach Summit knows your situation, tracks your progress across the year, and helps you build a case that's ready when the window opens.
