careerladder.fyi › Career paths › Sakana AI
How people got jobs at Sakana AI
What 52 people did before they joined Sakana AI: where they worked, what they studied, which role they came in through, and how long it took.
52
people in our data got there
71%
had a graduate degree first (of those listing one)
9
typical years since first job
- The most common step right before: from other companies (75%).
- The most common way in was Research (58% joined in that kind of role).
- Of those who list their education, 71% had a graduate degree before arriving and 29% a PhD.
- The typical person arrived about 9 years after their first job.
- The most common earlier employer: Google (12% had worked there).
How Sakana AI compares
- 29% had a PhD, above the 19% across other frontier AI labs.
- 10% came straight from Big Tech, below the 19% across other frontier AI labs.
- People arrived about 4 years earlier in their careers than at the typical one of the other frontier AI labs (9 vs 13 years since their first job).
Among people who list their education; places with at least 10 such people.
Show as a table
| People listing education | Share | |
|---|---|---|
| Google DeepMind | 58 | 41% |
| World Labs | 39 | 33% |
| Physical Intelligence | 31 | 32% |
| Sakana AI (this page) | 41 | 29% |
| Cohere | 40 | 28% |
| Hugging Face | 34 | 24% |
| Stability AI | 40 | 22% |
| SpaceXAI | 17 | 18% |
| Thinking Machines Lab | 29 | 17% |
| OpenAI | 41 | 17% |
| Anthropic | 28 | 14% |
| Black Forest Labs | 32 | 12% |
| EleutherAI | 26 | 12% |
| Reflection | 20 | 10% |
| Mistral AI | 36 | 8% |
| Skild AI | 39 | 8% |
| Poolside | 31 | 6% |
| Cognition | 27 | 4% |
| Perplexity | 24 | <1% |
Show as a table
| Median years | ||
|---|---|---|
| SpaceXAI | 18 yrs | |
| Anthropic | 15 yrs | |
| Mistral AI | 15 yrs | |
| OpenAI | 14 yrs | |
| Perplexity | 14 yrs | |
| Reflection | 14 yrs | |
| Thinking Machines Lab | 14 yrs | |
| Stability AI | 14 yrs | |
| Poolside | 13 yrs | |
| Cognition | 13 yrs | |
| Cohere | 13 yrs | |
| Black Forest Labs | 10 yrs | |
| Physical Intelligence | 10 yrs | |
| Google DeepMind | 10 yrs | |
| Skild AI | 9 yrs | |
| Sakana AI (this page) | 9 yrs | |
| World Labs | 8 yrs | |
| Hugging Face | 8 yrs | |
| EleutherAI | 5 yrs |
Which role they joined in
Kinds of role with fewer than 5 people are combined into "All other roles".
Show as a table
| People | Share | |
|---|---|---|
| Research | 30 | 58% |
| All other roles | 22 | 42% |
Show as a table
| People | Share | |
|---|---|---|
| Applied Research Engineer | 10 | 19% |
| Research Scientist | 10 | 19% |
What they did right before
Show as a table
| People | Share | |
|---|---|---|
| From other companies | 39 | 75% |
| From Big Tech | 5 | 10% |
| From banks & Fortune 500 | 5 | 10% |
| Other | 3 | 6% |
Where they worked before
A person can count under more than one employer.
Show as a table
| People | Share | |
|---|---|---|
| 6 | 12% | |
| Mercari, Inc. | 5 | 10% |
Education
Among the 41 of 52 people who list their education.
Show as a table
| People | Share | |
|---|---|---|
| Bachelor's | 12 | 29% |
| Master's | 16 | 39% |
| MBA | 1 | 2% |
| PhD | 12 | 29% |
A person can count under more than one school.
Show as a table
| People | Share | |
|---|---|---|
| University of Tokyo | 12 | 23% |
| Kyoto University | 5 | 10% |
How much experience they had
Among the 52 of 52 people with an earlier job listed.
Show as a table
| People | Share | |
|---|---|---|
| 1-3 years | 1 | 2% |
| 3-5 years | 5 | 10% |
| 5-8 years | 16 | 31% |
| 8-12 years | 11 | 21% |
| 12+ years | 19 | 37% |
Where would your path lead?
Paste your LinkedIn or resume and see the ladders people like you climbed to Sakana AI, and how each step changes your chances.
Plan my path to Sakana AI →
Paste your LinkedIn or resume and see the ladders people like you climbed to Sakana AI, and how each step changes your chances.
Plan my path to Sakana AI →
Related
How these numbers are made. From 3k+ public career histories collected for careerladder.fyi (dated jobs, degrees and publications). The collection was built around particular employers, so people who reached well-known tech, AI and finance employers are over-represented: read shares as "among people in this dataset", not as hiring odds. Names are shown only when at least 5 people are behind them; smaller groups are combined into "Other". No individual profile is shown. Updated November 2026.