Pick your lane
1.2 The Three Archetypes
As a Senior, you’re expected to be good at a wide range of tasks - metrics, experimentation, alignment, and more. As we discussed, you are a generalist. But to become Staff, you must be a company-wide expert in some domain. There are three broad archetypes in general:
- The Profile: Typically holds a PhD in Economics, Statistics, Causal Inference, or Physics.
- The Superpower: They solve the problems that standard libraries cannot solve. When the A/B test is contaminated by network effects, or when the data is observational and requires instrumental variables, you call them.
- The Value: They prevent the company from making confident decisions based on flawed math. They are the guardians of rigor.
- The Trap: Becoming a “solution looking for a problem.” They must learn to apply their heavy machinery only when the stakes ($$$) justify it.
- The Profile: Has spent 5+ years in a specific technical vertical (e.g., Search, Ads, Payments, Logistics). They understand the system at the same level as a Staff Backend Engineer.
- The Superpower: They know where the bodies are buried. They understand the hidden modes of failure, the data lineage dependencies, and the “ghosts” in the logging pipelines.
- The Value: Speed and risk mitigation. They can predict that a proposed experiment will break the upstream data warehouse before a single line of code is written.
- The Trap: Becoming a “historian.” They must use their knowledge to build the future architecture, not just protect the past.
- The Profile: A polymath who has worked across multiple domains (Growth, Core Product, Marketing). They are often the closest partner to the Director of Product or VP.
- The Superpower: Communication and synthesis. They can translate a p-value into a P&L statement. They are excellent at driving strategy (“What should we build?”) rather than just execution (“Did it work?”).
- The Value: Outcomes. They ensure the data team is working on the highest-leverage problems, not just the most interesting math problems.
- The Trap: Becoming a “shadow PM.” They must remember they are still Data Scientists—their authority comes from evidence, not intuition.
Takeaway: If you try to be all three at once, you fail. At Staff, you must spike.
The outer edges of this chart (5/5) offer exponential value: 10x more than being 4/5. This means you can be strategically weaker in certain areas than a generalist (2/5 instead of 3/5), as long as you're 5/5 in something else. A 5/5 score means you're one of the few people, maybe the only person, in a large company with that expertise.