Latin American PhDs
training frontier AI.
On your team in 14 days.

Vetted researchers for RLHF, SFT, DPO, and evals.
Coding-agent hires, inside your lab and on your hours.

Lifetime replacement · 97% retention · 325+ companies

Lifetime replacement
97% Retention
14 Days Time to Onboard

PhDs and Master's grads from leading Latin American universities

Universidade de São Paulo
Unicamp
Universidade Federal do Rio de Janeiro
Universidade Federal de Minas Gerais
Fundação Getulio Vargas
Universidad de Chile
Universidad de Buenos Aires
Universidad Nacional de La Plata
Universidad Nacional de Colombia

Domain expertise

The fields your models need

Every researcher has a PhD or Master’s in their field. Tell us the domain. We find the person.

  • Computer Science
  • Mathematics
  • Physics
  • Biology
  • Economics
  • Engineering
  • Linguistics and NLP
  • Law
  • Psychology
  • Chemistry

Talent preview

Researchers for frontier model work

Filter by the work your lab needs. Every profile is anonymized.

  • Camila R. portrait Bogotá, CO

    Camila R.

    Mathematics PhD · AI trainer

    • RLHF
    • Evals
    • SFT
  • Paula G. portrait Buenos Aires, AR

    Paula G.

    Chemistry PhD · Safety evaluator

    • Red-teaming
    • Evals
    • Annotation
  • Mariano F. portrait Santiago, CL

    Mariano F.

    ML engineer · Coding agents

    • Python
    • SFT
    • Evals
  • Marta S. portrait Quito, EC

    Marta S.

    Computational linguist · Annotation

    • Annotation
    • RLHF
    • Evals
  • Alodia P. portrait CDMX, MX

    Alodia P.

    Physics PhD · Reasoning evaluator

    • Evals
    • RLHF
    • SFT
  • Diana M. portrait Santiago, CL

    Diana M.

    Economics expert · Model evaluator

    • Evals
    • Annotation
    • RLHF
  • Marco L. portrait São Paulo, BR

    Marco L.

    Data engineer · Coding agents

    • Python
    • SFT
    • RLHF
  • Fausto H. portrait Quito, EC

    Fausto H.

    Law expert · Safety and alignment

    • Red-teaming
    • Evals
    • Annotation

Talent identities and images have been anonymized for privacy.

What they do

People for the work behind frontier models

They join your team and work inside your lab’s tools, process, and timezone.

Post-training researchers

Your researcher writes and grades RLHF, SFT, and DPO data inside your lab. They follow your rubrics and tools.

Model evaluation and safety

Your researcher runs model evals, red teams weak points, and checks annotation quality for safety and alignment.

Coding agents and reasoning

Your researcher builds RL environments and benchmarks for coding agents, reasoning, and domain work.

How we find researchers for frontier AI

The search starts with 637,000 candidates. You meet the few who fit the field and the work.

637,000 candidates, ready before you ask

Our database holds 637,000 candidates. Our pre-screened network has already cleared English, work history, and seniority. We narrow from there by field and the work your lab needs.

Map of the Americas Map of the Americas
  • Mexico City Mexico
  • Bogotá Colombia
  • Santiago Chile
  • São Paulo Brazil
  • Buenos Aires Argentina

8 layers before you interview

We screen 500+ candidates for each search across skills, English, communication, accent clarity, references, role fit, culture fit, and drive. You interview people who have already passed the bar.

Vetted researcher profile connected to eight numbered vetting layers Vetted researcher profile connected to eight numbered vetting layers

You set the final bar

Every placement clears our internal AI certification for prompt engineering and tool fluency. Then you interview the short list and test them the way your lab tests anyone.

Opus certification dashboard showing completed assessment modules

One rate. Everything handled.

One monthly rate covers placement, payroll, international compliance, and a lifetime replacement. You run the research inside your lab. We handle the rest.

Opus client dashboard showing active hires and monthly spend

Why Latin America

Better coverage for the same hiring budget

Researchers work your hours. Hiring costs 40 to 60% less than in the US or Europe.

40 to 60%

Lower cost

Hiring in Latin America costs 40 to 60% less than hiring in the US or Europe.

1 to 3 hours

US timezone overlap

Researchers work close to US Eastern time, so reviews and standups happen live.

Fluent

English and deep domain skill

We recruit published researchers from leading Latin American universities.

97% retention
14 days to onboard
325+ companies hired through Opus
500+ candidates screened per search

Questions

What AI labs ask before hiring

  • What is Opus AI training?

    We place vetted PhD and Master's graduates from Latin America inside AI labs and research teams. They work on RLHF, SFT, DPO, model evals, safety, alignment, coding agents, and annotation.

  • How does it work?

    You tell us the work and field. We search our database of 637,000 candidates, then screen 500+ people for the role. You interview the short list. Most hires start in 14 days.

  • Where are the researchers based?

    Across Latin America, including Brazil, Argentina, Mexico, and Chile. We recruit from leading universities across the region. Their working hours sit 1 to 3 hours from US Eastern time.

  • What tasks can they support?

    Your researchers write and grade RLHF, SFT, and DPO data. They build RL environments and benchmarks, run model evals, write rubrics, test coding agents, red team failures, and support safety and alignment.

  • Why hire AI researchers in Latin America?

    Hiring in Latin America costs 40 to 60% less than hiring in the US or Europe. Working hours line up with US teams. Researchers work in fluent English and come from the strongest universities in the region.

  • What does it cost?

    You pay one monthly rate. It includes payroll, international compliance, and support. Hiring in Latin America costs 40 to 60% less than hiring in the US or Europe.

  • How fast can someone start?

    Most placements start within 14 days. We search our database of 637,000 candidates and our pre-screened network, then screen 500+ people for each search.

  • What if it does not work out?

    Every hire comes with a lifetime replacement. If it does not work out, we replace them at no cost.

  • How are researchers vetted?

    Every researcher passes our 8-layer screen: domain skills, English, communication, accent clarity, references, role fit, culture fit, and drive. You meet people who have already passed the bar.

Meet your next AI researcher in 14 days.

Tell us the work and the field. We build the short list.

AI training talent

Tell us who you need.

Share the project and the domain. We will get back to you shortly.