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

Articul8 AI is seeking a Principal Research Scientist to define how we build, evaluate, and scale domain-specific models as a durable source of competitive advantage. You will lead research across the full model development lifecycle: domain data strategy, continued pre-training, supervised fine-tuning, post-training, evaluation methodology, and the strategic decisions that determine where Articul8 can create and sustain model superiority in the market.

• Set company-level technical direction for domain-specific model strategy — define how Articul8 builds, evaluates, scales, and sustains model superiority across continued pre-training, fine-tuning, post-training, and release quality standards, leveraging massively parallel agentic AI systems to compress strategic exploration cycles from months to days
• Architect the agentic model development paradigm for the organization — design the agent-orchestrated research infrastructure (experiment orchestration, data pipeline automation, continuous evaluation, competitive benchmarking) that enables every researcher at Articul8 to operate at a fundamentally higher level of depth, breadth, and velocity than would be possible alone
• Go deep: push the frontier of domain-specific model science — lead research on model adaptation methodology, data curation strategies, post-training methods (preference optimization, reward modeling, reasoning improvement, alignment), and training dynamics, deploying fleets of agentic systems to run exhaustive ablation studies, mixture experiments, and failure analyses in parallel
• Go broad: shape model strategy across all of Articul8's domains and verticals — define how the company identifies, prioritizes, and enters new model domains based on technical feasibility, customer value, and strategic differentiation, using agent-driven competitive intelligence and market analysis to scan the landscape continuously
• Define evaluation strategy as an agentic discipline — establish benchmark design, expert-grounded assessment, model failure analysis, and robustness standards, building always-on agentic evaluation harnesses that compare Articul8 models against leading open and closed alternatives and translate findings into concrete investment decisions in real time
• Lead cross-cutting research initiatives that multiply organizational capability — ensure advances in data perception, retrieval, post-training, and runtime orchestration strengthen the model layer, orchestrating parallel agent-driven research tracks across pillars so breakthroughs in one area compound across the platform
• Influence platform-level decisions — shape model lifecycle management, portfolio strategy, release criteria, and integration architecture, ensuring the platform is designed for humans and agentic systems to co-evolve and amplify each other