This video features Philip Su, who reached the highest level — Distinguished Engineer (IC9) — at both Meta and OpenAI, explaining the growth stages of engineers and what differentiates the top levels (E7, E8, E9). He emphasizes that what determines an engineer's level isn't raw technical skill but rather "scope of influence" and "time horizon." He also provides invaluable insights into the qualities and strategic thinking needed to reach the highest levels, where no clear checklist exists.
1. The Core Criteria for Engineer Levels: Scope of Influence and Responsibility
What's the most important criterion that separates engineer levels (E5, E7, E9, etc.)? Philip explains it comes down to a person's scope of influence and how large a project they can guarantee delivery on by themselves.
For example, an intern may struggle to guarantee completion of even a single feature. But a senior developer (E5–E6) should be able to lead and complete work for 10–15 people. So what about the higher E7 level?
For an E7, my data might be a bit dated, but I'd expect them to regularly influence the work of about 50 people. You'd be the person driving the technical direction for a team of roughly 50.
Product managers and engineering managers regularly consult you on key changes for the team. You're regarded as that influential.
You should also be able to personally deliver projects with about a 6-month scope. Meaning, if I give you a goal and say "I won't have time to talk for the next 6 months" and send you off into the coding wilderness, you should come back 6 months later with it done. That's the scale and scope of individual capability I expect.
As levels increase, both the time horizon of responsibility and the number of people influenced grow exponentially.
2. What Sets the Highest Levels (E8, E9) Apart: Quantitative and Qualitative Change
At the highest levels like E8 and E9, things get more complex. It's not just about managing more people (quantitative change) — qualitative change becomes crucial.
Philip says there are broadly two types of ways high-level ICs (individual contributors) prove their value:
- Deep specialists: Those with irreplaceable expertise in a specific domain.
- Broad generalists: "MacGyver" types who can handle a bit of everything on startups or new teams.
The important point is that the value of one senior-level engineer cannot simply be replaced by multiple junior-level engineers.
A great E8 engineer — someone with qualitative depth — can't simply be replaced by 4 E6 engineers and expect the same output. There's a level of work where no matter how many people two levels below you are thrown at it, you can't get the quality needed.
(...) In startup environments, generalists are extremely valuable. A great generalist E8 who can basically do a little bit of everything can't be surpassed by hiring three E6s.
But this qualitative difference is hard to define with HR's "promotion checklists" or item-by-item rubrics. This is precisely why many engineers get confused — they think they've checked every box but still don't get promoted.
3. The Ambiguity of E9 Promotion: Similar to Getting Tenure
Philip compares the E8-to-E9 promotion process to getting tenure at a university. It's not simply about publishing a certain number of papers — your fellow professors must recognize you as "professor-caliber" before you can become one.
In fact, when Philip was at Meta (then Facebook), the E9 level didn't even exist initially. It was only created when legendary developers like John Carmack joined, and the company needed a new level for them.
When the company was a few thousand people, if I remember correctly, there were only 3 engineers being considered for promotion to E9.
When the sample size is that small, it's impossible to create a checklist like "Do these 4 things as an E8, and we'll promote you to E9." Those 3 people were all different types of talent. They succeeded for different reasons.
(...) At level 8+ leadership tiers, I think a big part of it is "Do other E8s think you're an E8?" The reasons vary from person to person and are hard to articulate clearly. That's why it's very difficult to explain to people looking for a fixed promotion formula.
4. Growth Strategies Without Role Models
So what if you're in an environment (small company or weak team) without senior E8 or E9 engineers around? Philip acknowledges this as a "chicken-and-egg problem" and warns especially against being on "weak teams."
Even in a massive 140,000+ person company (during his Microsoft days), weak teams always exist. The problem with weak teams is that because your colleagues are mediocre, you start thinking you're exceptional.
It's like buying real estate. You never want to buy the most expensive house on the block. There's no room for the value to go up — you're already at the top. You don't want to be the worst house either. Being somewhere in the middle is best.
Without role models to learn from (the most expensive house), growth can stagnate. But Philip says something is even more important than finding a role model to imitate — having a coach who can give you accurate feedback.
At higher levels, I think having a great coach who can give you feedback matters more than finding someone to observe and emulate.
The person who has a say in your E8-to-E9 promotion needs to be able to tell you, "You're far from the qualitative bar we're looking for." Then you can have a meaningful discussion about "What actions do I need to take to reach that bar?" Getting qualitative input is far better than just finding someone to copy.
5. Real Example: The Decisive Project That Led to E9 Promotion
Finally, Philip shares his specific experience when he was promoted to E9. He grew Meta's London office engineering team from 12 people to 400–500 over 4–5 years.
Simply growing the team wasn't the reason for the promotion. It was solving complex problems that arose as the organization scaled and, above all, making strategic decisions by looking much further into the future than anyone else.
His specific example was about deciding intern hiring numbers.
One of my last meetings as site lead was convincing headquarters about how many interns we should hire in London that summer and how many full-time headcount to allocate.
I had already back-calculated the expected H-1B visa capacity for the interns who would convert to full-time the year after next. Do you follow? I was having conversations 18 months before things would actually happen, using data to determine what actions to take today for outcomes 18 months away.
Nobody else was thinking that way at the time. (...) Making decisions that affect a broader scope of people, or that require thinking on a much longer time horizon — that's what I believe was the primary reason for the promotion.
Conclusion
Growing into a top engineer goes far beyond just improving coding skills. It's about expanding your scope of influence, seeing a distant future (18+ months) that others can't, and creating irreplaceable qualitative value. There's no fixed answer or checklist, but finding a coach who can honestly assess your limits and cultivating strategic thinking are the keys to getting there.