AI Post-training's geography problem—and Latin America's accidental solution

The convergence of post-training demands, talent scarcity, and LatAm's research capacity is creating AI's next talent market.

Post-training work—the phase where AI models are refined into reliable, domain-specific tools—requires a specific kind of talent that's increasingly hard to find in the U.S.

It demands PhD-level domain expertise across fields like medicine, physics, engineering, computer science, and more. It requires researchers who can evaluate model outputs, design safety frameworks, and adapt systems for specialized applications. And it needs to scale to thousands of expert hours per model.

Demand is outpacing domestic PhD supply by orders of magnitude. There’s one region that can help close the gap—and it remains largely untapped.

The talent shortage by the numbers

The AI industry has a new bottleneck, and it's not compute or data. It's people with PhDs.

Yet, U.S. universities awarded fewer than 3,000 Computer Science PhDs in 2023, according to the National Science Foundation, with Physics and Mathematics adding roughly 4,000 more. And nearly half are international students, many of whom face visa constraints after graduation.

As AI systems move from pre-training to post-training, the work fundamentally changes. Pre-training is capital-intensive: massive GPU clusters running overnight. Post-training is labor-intensive: thousands of hours of expert evaluation, safety testing, and domain adaptation by people who understand the domains where AI is being deployed.

McKinsey research shows the number of U.S. workers in roles requiring explicit AI fluency grew sevenfold between 2023 and 2025—from 1 million to 7 million.

LinkedIn's 2026 Jobs on the Rise report makes the shift concrete: AI Engineer, AI Consultant, and AI/ML Researcher now rank among the five fastest-growing roles in the United States. Also in the top five: Data Annotators, the frontline workers who label and review data to train AI models.

The math doesn't work. Demand is growing faster than supply.

Why Latin America specifically

While AI labs compete for the same constrained pool of U.S. researchers, Latin America graduates tens of thousands of PhDs annually across all disciplines, with thousands in STEM fields like computer science, mathematics, physics, and engineering.

Sources: CAPES (Brazil), CONACYT (Mexico), CONICET (Argentina), CNED (Chile), 2023/2024

These aren't insular research programs. Researchers from institutions such as USP, Unicamp, UNAM, and CONICET regularly collaborate with peers at MIT, Stanford, UC Berkeley, Oxford, and Max Planck, publishing jointly in venues like Nature, IEEE, and ACM.

The combination is what matters: research quality that is competitive with U.S. universities, costs 40-60% lower than domestic hiring, and time zones that enable real-time collaboration when it's needed.

Post-training involves constant iteration. A domain expert identifies an edge case, an engineer adjusts the approach, the expert validates the output. This loop repeats hundreds of times per model. Brazil, Mexico, and Argentina operate within 1 to 4 hours of U.S. Eastern Time. A researcher in São Paulo starts their workday as New York wakes up. A team in Buenos Aires overlaps 4 hours a day with San Francisco.

When problems can be resolved in hours instead of days, development velocity compounds. It's not the primary reason to hire in Latin America, but it's a meaningful advantage that becomes more valuable at scale.

Why this moment is different

For years, the AI talent conversation focused on machine learning engineers building foundation models. That's capital-intensive work concentrated in a handful of frontier labs.

Post-training changed the equation. It's where models become useful—fine-tuned for radiology, contract law, financial compliance. The work requires domain experts who understand clinical reasoning, case law, or regulatory frameworks, combined with strong quantitative training.

Three forces are converging right now:

Demand: Post-training has become the primary bottleneck in AI deployment, often requiring PhD-level expertise the U.S. can't supply at scale.

Supply: Latin America produces researchers well aligned with the domain and quantitative backgrounds post-training demands—and most can't access frontier AI work locally.

Infrastructure: Remote work, normalized during the pandemic, removed the geographic constraints that once required researchers to relocate to contribute to cutting-edge AI development.

The result is a market mismatch being solved by geography.

What this means for AI labs

Forward-thinking organizations are starting to explore a different model: distributed research teams that combine U.S. leadership with Latin American depth.

The AI industry spent two years learning that compute alone doesn't solve intelligence. The next lesson is becoming clear: intelligence without distributed expertise doesn't solve real-world problems.

Post-training demands are growing faster than U.S. PhD production. Latin American universities are producing thousands of qualified researchers annually. Time zones enable real-time collaboration. Economics make it sustainable.

The structural forces are in place. What's missing is execution.

For AI labs, research-driven companies, and organizations deploying specialized models, the question is no longer whether to build distributed teams. It's whether to move while the talent market is still underutilized—or wait until the competition figures it out.

How to access this talent pool

Athyna is a global talent platform with years of experience helping startups and large enterprises build high-performing global teams in Latin America. Long before AI training became a mainstream challenge, Athyna was already developing the infrastructure needed to source, vet, and deploy highly skilled professionals internationally, supported by an AI-powered matching platform.

That experience is now focused on one of the most acute bottlenecks in AI: access to PhD- and Master’s-level expertise for post-training, evaluation, and domain-specific work.

Athyna Intelligence connects companies with advanced researchers across Brazil, Mexico, Argentina, and beyond, pre-vetted not only for technical depth, but also for English fluency, research experience, and readiness to collaborate with global AI teams. The result is faster access to scarce expertise, U.S.-aligned collaboration, and the ability to scale research capacity at 40–60% lower cost than comparable U.S. hires.

As AI systems grow more complex, the ability to tap into global pools of scientific talent will increasingly determine which organizations can move fastest and most responsibly. The next breakthrough in AI may not come from Silicon Valley alone. It may come from collaborations with São Paulo, Buenos Aires, or Mexico City. The only question is: Will your team be part of it?

—

*Sponsored by Athyna. We have equity in the company.

Disclaimer: The Bay Area Times is a news publisher. All statements and expressions herein are the sole opinions of the authors or paid advertisers. The information, tools, and material presented are provided for informational purposes only, are not financial advice, and are not to be used or considered as an offer to buy or sell securities; and the publisher does not guarantee their accuracy or reliability. You should do your own research and consult an independent financial adviser before making any investments. Neither the publisher nor any of its affiliates accepts any liability whatsoever for any direct or consequential loss howsoever arising, directly or indirectly, from any use of the information contained herein. Assets mentioned may be owned by members of the Bay Area Times team.

Please read our Terms of Service and our Privacy Policy before using Our Service.