Skip to content
The Growth Track
← All articles
  • Career ladders
  • Engineering management

How to build a software engineering career ladder

A practical, step-by-step guide to defining tracks, levels, areas and expectations for engineers, without writing a 40-page document nobody reads.

Rachid Baaziz

· 5 min read

A career ladder answers two questions every engineer eventually asks: what is expected of me now? and what do I need to show to get to the next level? When nobody can answer them clearly, promotions feel political and reviews feel arbitrary, even when every manager acts in good faith.

The good news: a useful ladder is much smaller than most people expect. This guide walks through the structure we recommend, the decisions you need to make, and the mistakes that make ladders fail.

Start with the decisions, not the document

Before writing a single expectation, settle four structural questions. They shape everything else.

  1. Which tracks do you need? Most engineering organizations need an individual contributor (IC) track and a management track. Engineers should be able to grow without becoming managers.
  2. How many levels per track? Fewer than you think. Five IC levels (junior to principal) and three management levels cover most companies up to a few hundred engineers.
  3. Which areas matter? Group what you care about into three to five areas, such as Execution, Technical Mastery, Collaboration and People Leadership.
  4. Who owns the ladder? One accountable owner (often the CTO or VP Engineering) keeps it coherent. Engineering managers contribute; they don't each maintain their own version.

Write these decisions down before anything else. They are the skeleton every expectation hangs on.

Define tracks and levels

A track is a path; a level is a step on it. Give each level a short, stable name, and avoid tying levels to job titles that HR may change for other reasons.

A common structure for a software engineering organization looks like this:

TrackLevels
Individual ContributorL1 Junior Engineer → L2 Software Engineer → L3 Senior Engineer → L4 Staff Engineer → L5 Principal Engineer
Engineering ManagementM1 Engineering Manager → M2 Senior Engineering Manager → M3 Director of Engineering

Two rules keep levels meaningful:

  • Each level must be distinguishable from its neighbours. If you can't explain the difference between L3 and L4 in one sentence, you have one level too many.
  • Levels describe scope and impact, not years of experience. Tenure is a poor proxy for the work someone actually does.

Choose your areas, and break them down only when needed

An area is a broad part of performance or growth you want to define, such as Ownership or Technical Excellence. Many areas work best described as a whole, with one or two sentences per level.

When an area is too broad to describe in a sentence or two, break it into capabilities: the specific aspects within it. Execution might hold Delivery, Ownership and Quality; Technical Mastery might hold Architecture, Engineering Practices and Systems Thinking.

Keep the list short. Every capability you add multiplies the number of expectations you have to write and the number of things managers have to assess. Start simple, and add capabilities only where an area needs more detail.

Some areas apply to only one track. People Leadership belongs on the management track; Technical Mastery may apply only to individual contributors. Make that explicit rather than expecting managers to ignore what doesn't fit.

Write expectations people can recognize

This is where most ladders succeed or fail. An expectation describes what good looks like at one level, for an area or one of its capabilities.

Weak expectations are vague and interchangeable:

Demonstrates strong ownership.

Strong expectations are concrete enough that two managers would agree on whether someone meets them:

Owns a service or product area and proactively improves its health.

A few guidelines help:

  • Describe observable behaviour and outcomes. "Breaks down ambiguous projects into milestones and delivers them predictably" can be checked against real work.
  • Make scope grow between levels. The same area often reads as task → feature → system → several teams → organization as you climb.
  • Write one or two sentences per area and level. If an expectation needs a paragraph, it is probably several expectations, or a sign the area deserves capabilities.
  • Avoid adjectives that only rank. "Excellent", "strong" and "advanced" tell people they should be better without saying how.

Decide how you will evaluate against it

A ladder is only useful if reviews use it. Decide early:

  • What is rated: each area (or capability), each expectation, or nothing at all (written feedback only). Rating every expectation gives the most detail but the most work.
  • The rating scale: four or five clearly named points (for example Developing, Meets expectations, Exceeds expectations) work better than numbers alone.
  • Whether reviewers attach evidence: links to design documents, incidents or shipped work keep reviews grounded in facts.
  • Whether to compute an overall score, and if some areas should weigh more at some levels.

These choices belong in the framework itself, so every manager runs reviews the same way.

Roll it out in small steps

You don't need a perfect ladder on day one. A rollout that works:

  1. Draft the structure and two levels in detail.
  2. Review the draft with a few senior engineers and managers. Ask them where they would place real people, and why.
  3. Fill in the remaining levels.
  4. Place every engineer on a track and level, and share the ladder with the whole team.
  5. Use it for goals during the year and for the next review cycle.
  6. Revise it after each cycle. Treat it as a product with versions, not a policy carved in stone.

Common mistakes

  • Copying a big-tech ladder as is. Their levels reflect their scale. Use public ladders as inspiration, then adapt.
  • Too many levels. More levels mean more promotion debates, not more clarity.
  • Ladders that only exist in a document. If goals and reviews don't reference the ladder, people stop reading it.
  • Changing the ladder in the middle of a review cycle. Freeze the version a cycle uses, and introduce changes for the next one.

A clear ladder won't make every career conversation easy. It will make them about the same things, for everyone.