AI research scientist

AI Research Scientist

Writer is an enterprise AI platform that powers marketing and revenue teams at Fortune 500 companies, including three of the top five US banks, two of the five largest global CPG conglomerates, and four of the top five US healthcare insurers. Founded in 2020 and backed by major investors including Insight Partners, ICONIQ, Radical Ventures, and Salesforce Ventures, Writer develops autonomous agents and large language models that help organizations scale their go-to-market operations. The company combines frontier AI capabilities with enterprise governance and compliance infrastructure designed for mission-critical deployments.

Writer's AI research function focuses on building the scientific foundation for enterprise-grade AI systems that serve hundreds of thousands of users daily. This role offers a rare opportunity to conduct research that simultaneously advances the field and ships directly into production systems at scale. You'll work at the intersection of academic rigor and real-world impact, driving improvements to language models and agentic systems that thousands of enterprises rely on.

What you'll do

• Lead research projects spanning data synthesis, novel model architectures, model training, evaluation, and production deployment as a core member of the research team

• Design and execute post-training experiments using supervised fine-tuning, reinforcement learning from human feedback, RLAIF, DPO, GRPO, and emerging alignment techniques, with emphasis on improving multi-step reasoning, planning, and tool use in enterprise agentic workflows

• Create novel evaluation benchmarks and methodologies that measure how well models perform on complex, real-world enterprise tasks beyond existing limitations

• Build scalable data synthesis and curation pipelines generating high-quality training signals, including LLM-as-judge frameworks, synthetic data generation, and adversarial dataset construction

• Translate research insights into concrete improvements to Writer's enterprise-grade language models by shaping model architecture and training roadmap alongside research engineering and product teams

• Publish original research at top-tier venues including NeurIPS, ICLR, ICML, ACL, and EMNLP, representing Writer at the frontier of AI research

• Establish high standards for scientific rigor and experimental design within the research team through close collaboration with fellow researchers and engineers

What you'll bring

• Four or more years of hands-on ML research experience, including deep expertise in large language model pre-training and post-training—you have scaled model training, debugged distributed systems, and shipped improvements with measurable impact

• Ph.D. in Computer Science, Machine Learning, NLP, or related field, or equivalent demonstrated research experience backed by a strong portfolio of independent published work

• Expert knowledge of post-training methods including SFT, RLHF, RLAIF, DPO, and GRPO, with proven track record applying these techniques to production-grade systems

• Strong proficiency in Python and PyTorch or JAX, with engineering depth to independently build and scale training pipelines, evaluation infrastructure, and data synthesis workflows

• One or more publications at competitive ML/AI venues such as NeurIPS, ICLR, ICML, ACL, or EMNLP demonstrating ability to originate ideas and execute multi-month research agendas independently

• Hands-on experience designing or evaluating agentic systems that plan, reason through multi-step tasks, use tools, and recover from errors, with nuanced understanding of failure modes and solutions

• Collaborative approach across research, engineering, and product with ability to communicate complex technical concepts clearly to diverse audiences

• Ownership mindset driving projects end-to-end with urgency and accountability for real customer impact

What they offer

• Generous paid time off plus company holidays

• Medical, dental, and vision coverage for you and your family

This is a hybrid role based in San Francisco, reporting to the VP of AI Research. Writer is open to hiring at multiple experience levels with compensation scaled accordingly.