The opportunity
As a Research Manager within Datadog AI Research (DAIR), you will lead a team working on fundamental research in foundation models and world models while remaining deeply engaged in the technical direction of the work. You will advance large-scale pre-training and multimodal…
What you'll do
Lead research in foundation models, world models, and multimodal learning,: setting the technical direction for ambitious research programs grounded in observability and security
Guide and mentor researchers and research engineers while remaining closely: involved in research strategy, experimentation, model development, and technical problem-solving
Train large-scale multimodal models on diverse telemetry data, including: metrics, logs, traces, topology, events, and other non-text modalities
Advance approaches to pre-training, representation learning, world modeling,: scaling, and evaluation for models that learn the dynamics of complex distributed systems
Collaborate with cross-functional teams across Research, Product, and: Engineering to translate research advances into scalable Datadog capabilities
Contribute to research publications, present at top-tier conferences such as: NeurIPS, ICLR, and ICML, and help open-source key model artifacts and benchmarks
What they're looking for
- You hold a PhD in Computer Science, Machine Learning, or a related field, or: have equivalent experience, with deep expertise in areas such as foundation models, world models, multimodal learning, or generative modeling
- You have demonstrated the ability to lead technically ambitious research at: meaningful scale, whether in an industry research lab, startup, academic environment, or another research setting
- You have extensive hands-on experience designing, training, or evaluating: large-scale deep learning models (such as large language models), with experience in multimodal or non-text data considered a strong plus
- You have a track record of research impact through influential publications,: significant model or system contributions, widely used research artifacts, or equivalent technical achievements
