AI Jobs in Brazil in 2026: Nubank ML, Mercado Libre AI, and the Portuguese-Language LLM Wave

Brazil's AI sector grew through 2024-2025 from a small data-science cluster into a serious applied AI market. Here's where the real AI engineering and research roles are — at Nubank ML, Mercado Libre, MaritacaAI, and the emerging Portuguese-LLM ecosystem.

Brazil's AI Moment Is Real, If Specific

Brazilian AI hiring through 2022 looked like everywhere else: lots of "data scientist" roles that were really analytics-with-pandas jobs, a small applied-research community at the major universities (USP, UNICAMP, UFRJ), and one or two startups that overclaimed about their ML stacks. The 2024-2025 shift was tangible. Nubank's ML organisation crossed several hundred ML engineers and applied scientists. Mercado Libre's São Paulo and Buenos Aires ML teams started publishing at top conferences. MaritacaAI launched as a credible Portuguese-language LLM lab. And serious AI infrastructure work (Latitude.sh, edge-AI startups, fintech ML at Stone and PagSeguro) became a real career path.

The market is smaller than Singapore or Tel Aviv but it's the largest English-friendly AI hiring market in Latin America. For international AI engineers considering the Western Hemisphere outside the US, Brazil belongs on the shortlist in a way it didn't three years ago.

Where the AI Roles Actually Sit

Nubank AI / ML

The largest single concentration of ML engineering in Latin America. Several hundred ML engineers and applied scientists across credit underwriting, fraud detection, customer service AI (including a deployed conversational AI handling significant call-centre volume), and personalised banking products. Strong publication record at top conferences (KDD, NeurIPS workshops). Compensation at senior+ levels: R$30,000-80,000/month + meaningful Nubank equity at NYSE-listed shares. Hires both PhD researchers and strong production-ML engineers; bar is FAANG-level.

Mercado Libre ML (São Paulo + Buenos Aires)

Distributed ML organisation across multiple countries. São Paulo focus on Brazilian-specific personalisation, search relevance for Portuguese, and fraud detection for the local payments side (Mercado Pago). Strong on applied recommendation, search, and computer vision for product imagery. Less concentrated on frontier research than Nubank but more breadth across applied problems.

iFood ML

Logistics-heavy applied ML. Courier-restaurant-consumer matching, ETA prediction at scale, demand forecasting, restaurant recommendation, fraud detection. Strong production-ML focus rather than research-oriented. Pay is competitive (R$18,000-50,000/month at senior+ levels) with strong equity through Prosus.

Stone, PagSeguro AI

Credit-risk ML, payments fraud detection, applied NLP for customer service. Less prestigious than Nubank but real production-ML work at meaningful scale. Reasonable career path if Nubank is closed for the year.

Big Tech LATAM AI Teams

Google Brasil AI / Google Cloud LATAM — Cloud AI specialist roles, applied research collaborations with USP and UNICAMP, customer engineering for AI deployments. Smaller than Singapore or Madrid offices but established
Meta São Paulo AI — small but growing, focused on Portuguese-language NLP and integrity work
Microsoft Brasil AI — Copilot for Brazilian market, AI specialists for major Brazilian customers, applied research
Amazon AWS Brasil ML — applied ML for AWS AI services and customer engineering across LATAM

The Emerging Portuguese-Language LLM Ecosystem

MaritacaAI (São Paulo)

The most credible Brazilian LLM lab in 2026. Founded in 2023 by University of São Paulo (USP) researchers focused on Portuguese-language LLMs. Released the Sabiá and Maritaca models trained primarily on Brazilian Portuguese data. Series A funded in 2024. Hires research engineers, applied AI engineers, and infrastructure engineers; Portuguese fluency strongly preferred for research-heavy roles but English-friendly at the engineering level.

Pluri / 23Tech (São Paulo)

Smaller Brazilian AI startups working on enterprise applications of LLMs for Portuguese-speaking markets. Less visible internationally but interesting for engineers who want early-stage equity in the regional AI build-out.

USP, UNICAMP, UFRJ Research Centres

The traditional Brazilian AI research universities. Strong on machine learning theory, applied NLP, and computer vision. Academic-tier salaries (R$8,000-18,000/month for postdoc and research engineer roles) but world-class research environment, and a real pipeline into Nubank, Mercado Libre, and Big Tech LATAM after academic stints.

What These Roles Pay in 2026

Monthly gross compensation in BRL (annual conversion in parentheses):

Junior ML Engineer (0-2 yrs): R$8,000 – R$13,000/month (R$96K – R$156K/year)
Mid ML Engineer (3-5 yrs): R$13,000 – R$22,000/month
Senior ML Engineer (5-8 yrs): R$22,000 – R$38,000/month
Staff ML Engineer: R$38,000 – R$65,000/month
AI Research Scientist (PhD, mid): R$25,000 – R$45,000/month at corporate labs
Nubank senior+ ML: R$35,000 – R$80,000/month + equity
Big Tech LATAM senior+ AI: R$50,000 – R$90,000/month + equity, total comp closes 65-80% of US Big Tech
MaritacaAI / startup AI roles: R$20,000 – R$50,000/month + early-stage equity

What's Different About Brazilian AI Hiring

Portuguese-language NLP is a real specialisation. Models trained on Brazilian Portuguese, working through the language's specific grammatical structures, handling Portuguese honorifics in conversational AI — all genuine specialisations. Engineers who can credibly work on this stand out
Production-ML experience matters more than papers. Nubank, Mercado Libre, and iFood ML teams will pick a candidate with strong production scars over one with NeurIPS papers but no production exposure, more often than US labs would
The academia-industry pipeline is functional. USP, UNICAMP, and UFRJ regularly send PhD students to Nubank, Mercado Libre, and Big Tech LATAM. Maintaining academic collaborations is part of the senior-AI-engineer culture in São Paulo in a way it isn't in most US tech companies
Multilingual ML focus. Working in Portuguese + Spanish + English simultaneously is common (Mercado Libre serves all three) and creates a real differentiator for engineers who can work cross-lingually

How to Apply, Practically

Direct careers sites for Nubank ML, Mercado Libre AI, iFood, MaritacaAI — all run their own pipelines
LinkedIn Brasil with "Machine Learning" and "AI" filters set to São Paulo, Rio, or remote-Brazil
Gupy for broader Brazilian tech and AI hiring
USP, UNICAMP, UFRJ research groups directly for academic and research-engineer positions
SBIA (Brazilian Symposium on Artificial Intelligence) — annual conference where Brazilian AI labs actively recruit

A Specific Contrarian Move

Most international AI engineers evaluating Latin America in 2026 default to "join Nubank ML." It's the right answer for most people and the path I'd typically recommend. Here's the contrarian alternative worth considering: MaritacaAI is the most credible non-US, non-Chinese Portuguese-language LLM lab being built. The first 30 engineers and researchers there will have outsized impact on whether Brazil develops sovereign LLM capability. The compensation is below Nubank but the equity stake and the optionality of being early in a frontier-AI bet aimed at a 250-million-speaker language market is unusual. If you believe Latin America deserves AI infrastructure that isn't a US or Chinese model translated into Portuguese, MaritacaAI is the most interesting career bet in 2026. Most people will pick Nubank and not be wrong. A few people who pick MaritacaAI could end up shaping how 250 million people experience AI for the next decade.