Principal Research Scientist (AI/ML)

Location: Brazil, hybrid

Company

hirify.global develops enterprise GenAI platforms combining domain-specific models, agentic reasoning, model evaluation, and multimodal understanding for regulated industries.

What you will do

• Set company-level technical direction for domain-specific model strategy, including pre-training, fine-tuning, post-training, evaluation, and release standards.
• Architect agent-orchestrated research infrastructure for experiment orchestration, data pipeline automation, continuous evaluation, and competitive benchmarking.
• Lead research on model adaptation, data curation, preference optimization, reward modeling, reasoning improvement, alignment, and training dynamics.
• Define evaluation strategy through benchmark design, expert-grounded assessment, failure analysis, and robustness standards.
• Shape model lifecycle management, portfolio strategy, release criteria, and platform integration architecture across energy, semiconductor, finance, aerospace, and supply chain domains.
• Mentor Staff and Senior researchers while maintaining hands-on contributions through technical work, publications, patents, and externally visible research.

Requirements

• PhD or MSc in Computer Science, Machine Learning, NLP, or a related field.
• 10+ years of AI/ML research experience, including 4+ years developing LLM-based systems, with models or systems deployed in production.
• Hands-on mastery of continued pre-training, supervised fine-tuning, post-training alignment, and production evaluation.
• Experience designing expert-grounded evaluation, systematic error analysis, robustness testing under distribution shift, and high-stakes deployment criteria.
• Experience training or adapting models on large GPU clusters with distributed frameworks such as DeepSpeed, FSDP, or Megatron-LM.
• Proficiency in Python and PyTorch, along with organizational research leadership and strategic decision-making experience.

Nice to have

• Experience building domain-specialized models that outperform general-purpose alternatives on measurable tasks.
• Hands-on experience with RLHF, DPO, reward modeling, or constitutional approaches.
• Deep experience in deduplication, data mixture design, and quality scoring for model development.
• Publication record at top-tier venues and experience taking research from prototype to production.
• Domain expertise in energy, semiconductors, finance, aerospace, or supply chain.

Culture & Benefits

• Research is focused on accurate, explainable, auditable, and trustworthy AI for high-stakes environments.
• Work combines research, engineering, product, and domain expertise.
• Professional principles emphasize humility, evidence-based decisions, measurable outcomes, care for colleagues, ambitious research bets, and responsible AI.