Principal Applied AI Researcher - Domain-Specific Models (Brazil)

About the role

Articul8, a GenAI platform serving regulated industries such as energy, semiconductor, finance, aerospace and supply chain, is looking for a Principal Research Scientist in Brazil to define how the company builds, evaluates and scales domain-specific models. You will lead research across the full model development lifecycle, from data strategy and continued pre-training through fine-tuning, post-training, evaluation and release quality.

Key responsibilities

• Set company-level technical direction for domain-specific model strategy, using massively parallel agentic AI systems to shorten strategic exploration cycles.
• Design the agent-orchestrated research infrastructure (experiment orchestration, data pipeline automation, continuous evaluation, competitive benchmarking).
• Lead research on model adaptation, data curation, post-training methods (preference optimization, reward modeling, reasoning improvement, alignment) and training dynamics.
• Shape model strategy across all domains and verticals, using agent-driven competitive intelligence and market analysis.
• Establish evaluation strategy: benchmark design, expert-grounded assessment, failure analysis and robustness standards, with real-time comparison against open and closed alternatives.
• Orchestrate parallel research tracks on data perception, retrieval, post-training and runtime orchestration.
• Influence model lifecycle management, portfolio strategy, release criteria and integration architecture.
• Mentor staff and senior researchers, coach on agent-augmented research programs and help shape hiring.
• Maintain a personal research contribution through technical work, publications and patents.

Required profile

• PhD or MSc in Computer Science, Machine Learning, NLP or a related field.
• 10+ years in AI/ML research with a track record of production models that changed outcomes, and 4+ years developing LLM-based systems.
• Hands-on experience across the full model lifecycle, including continued pretraining, fine-tuning, post-training alignment and production evaluation.
• Experience designing evaluation beyond leaderboard metrics, and training/adapting models on large GPU clusters with distributed frameworks.
• Strategic leadership: shaping research direction and allocating research investment.

Required skills

• Proficiency in Python and PyTorch.
• Preferred: RLHF, DPO, reward modeling, constitutional approaches, data curation (deduplication, mixture design, quality scoring), and domain expertise in energy, semiconductor, finance, aerospace or supply chain.