Senior Research Scientist

Dataminr•Arlington, VA

1d•$190,000 - $245,000•Hybrid

About The Position

As a Senior Research Scientist at Dataminr, you will work on AI problems in real-time threat and risk intelligence, advancing large language models, generative AI, and agentic systems and bringing research into products. You will train and adapt LLMs, build multimodal and agentic systems that reason over live global data, and develop approaches to surface important insights. This is a hybrid position based in Dataminr’s New York City or Arlington, Virginia office. The role focuses on building and deploying AI models that identify real events from fast-moving, multilingual data; advancing multimodal and LLM research; and taking research from experimentation through production deployment. You will collaborate with engineers, product experts, and domain specialists, and contribute to the research community through publications, program committee service, and AI for Good collaborations with nonprofit organizations.

Requirements

Ph.D. in Computer Science or a related field.
Research depth in one or more of: LLM training and adaptation, distillation, evaluation and benchmarking, agentic systems, retrieval and grounding, search, recommendation, multimodal models, or temporal and streaming reasoning.
Research track record demonstrated through publications at venues such as NeurIPS, ICML, ICLR, ACL, AAAI, COLM, or CVPR.
Hands-on experience training and adapting models at scale.
Strong Python and PyTorch fundamentals.
Comfort with modern training and inference tools such as vLLM and Langfuse.
Ability to independently design experiments, choose principled metrics, and demonstrate progress.
Fluency with AI coding and research agents and experience using them to accelerate work.
Ability to own research projects from initial idea to shipped result.
Ability to provide cross-functional technical guidance and mentorship.

Responsibilities

Build and deploy AI models that determine in seconds whether bursts of posts across 150 languages signal real events.
Advance multimodal and LLM research, including systems that verify video from unfolding crises and agents that keep event briefs accurate as facts change.
Take research from experiments into production and ship models in Dataminr products.
Develop AI systems that reason over live global data at massive scale.
Partner with engineers, product experts, and domain specialists to deliver value for existing and new customers.
Share research through publications at top-tier venues, program committee service, and AI for Good collaborations with nonprofit organizations.
Own research projects from initial idea through shipped result and bring teammates along.
Provide cross-functional technical guidance and mentorship to help other scientists grow.