AI hiring has shifted toward work experience and practical output, so companies should soften academic requirements and assess what candidates have built.
2
San Francisco, Seattle, and New York remain the main AI talent markets, and recruiting teams should track where talent and funding are moving.
3
Recruiting works better when companies use retention, career history, timing, and a specific company narrative instead of relying mainly on salary and equity.
Summary
Heath Black presents SignalFire's approach to recruiting AI teams using data from its Beacon platform. He argues that hiring filters need to reflect the market's shift away from degrees and toward practical experience, especially as more AI work involves products, ML operations, and software. Location still matters: San Francisco remains the largest AI talent market, with Seattle and New York also concentrating talent. Timing matters because retention, career stage, generation, and company history affect whether someone may move and whether they fit a startup's current stage. Black also argues that salary and equity have become weaker recruiting tools. AI engineers command a premium, while fewer employees exercise vested shares. Companies therefore need a more specific narrative about their mission, team, pace, learning opportunities, market, and past progress. The talk gives hiring teams a data-based framework for filtering, locating, timing, and persuading candidates.
Practical work now matters more than academic credentials for many AI roles
Black says AI startups have hired fewer engineers from top schools and fewer engineers with PhDs than they did in 2015. The share of hires from top schools fell from 27% to 15% by 2023, while the share with PhDs fell from 16% to 7%. Research scientists are more likely to have advanced degrees, but Black says they still make up less than half of people in those roles. He connects the change to the market moving from foundational ML research toward applied work such as MLOps, product development, software, and understanding users. He advises companies to assess open-source contributions and things candidates built outside class, and to remove or soften academic requirements in job postings.
Employee movement reveals which companies are producing the talent startups want
Black describes a shift in AI talent from large technology companies such as Google, Uber, Meta, and Apple toward what SignalFire calls the AI League. He says companies need to study both where people go and where they came from. SignalFire's employee-flow data shows different patterns between companies, including positive movement from DeepMind to OpenAI and a negative trend for Cohere in the example he presents. This information can shape recruiting filters and help teams understand which backgrounds are relevant for a role. Work experience now carries more weight than education, including for newer workers whose evidence may come from open-source projects or independent builds.
AI talent remains concentrated in a few cities despite distributed work
Black rejects the idea that San Francisco is dead as a startup or AI hiring market. He says San Francisco accounts for about 29% of startup engineers, down from 33% in 2013 but rising again since 2021. New York and Seattle have both doubled their share of startup engineers over the same period. For AI specifically, San Francisco has about 35% of AI engineers, Seattle about 22%, and New York about 10%. He also connects San Francisco's talent concentration to funding, saying it receives nearly 38% of early-stage funding for AI startups while accounting for 26% of early-stage US funding overall. Recruiters should watch both talent and capital flows.
Recruiting timing depends on retention and a candidate's fit with the company stage
Black treats timing as two related questions: when someone is most likely to leave, and whether they are willing to join a company at its current stage. He compares retention across AI League companies, citing about 66% four-year retention at Anthropic and about 43% at Perplexity. Retention data can help create what he calls a poachability score, or an estimate of how likely someone is to answer outreach. He also discusses differences between generations. In 2023, nearly 27% of Gen Z left their job, more than twice the rate he gives for Gen X. Within four years of graduating, Gen Z had about 2.2 jobs compared with 1.1 for Gen X.
A company's historical composition helps identify candidates who fit its next stage
SignalFire's historical composition tool shows the structure of companies at different points in their histories. Black gives examples such as identifying the sales leaders who moved a company from $1 million to $10 million, or the first three engineers who helped ship a product a startup wants to compete with. This history can help recruiters understand a person's risk profile, motivations, and likely fit for the company's current stage. It can also help identify people who may have the experience to produce a tenfold increase in impact as the company grows. Black recommends tracking when people join and leave admired companies and watching for profile changes that may signal a move.
A recruiting story needs to explain the company's arc
Black uses Kurt Vonnegut's shapes of stories to explain why recruiting narratives need an arc. A company does not need to describe itself as being in despair, but it should explain its past wins, why it is where it is, and where it is going. Salary and equity used to carry much of the recruiting pitch, but Black says those tools have weakened. From November 2022 to November 2024, he cites a 1.6% increase in average tech salary and a decline in equity grants. AI engineers receive a 5% salary premium and a 10% to 20% equity premium over other engineering roles, making compensation alone harder for startups to use.
Teams need reasons to join that go beyond compensation
Black says equity is also less persuasive when candidates are uncertain about company valuations, the cost of exercising shares, and changing markets. He cites a fall in the share of people exercising vested stock from 55% to 33% between an earlier period and Q2 2024. Companies therefore need to describe what candidates will experience and accomplish. His examples include close work with founders, collaborative teams, speed, low friction, a large mission, career growth, expanding markets, and complex problems. The recruiting narrative should be specific to the company rather than relying entirely on money and equity.
"The reality is experience and what you're building matters more than where you get a degree."07:28
Who should watch
You are hiring AI engineers and need to decide whether degrees, research credentials, or practical work should drive your initial filter.
Your recruiting team is spread across locations and wants evidence about where AI talent and funding are actually concentrated.
You are competing for expensive candidates and need a company story that gives them reasons to join beyond salary and equity.
