Staff Research Scientist - Post-Training for Observability Native Models
Staff Research Scientist - Post-Training for Observability Native Models
What You'll Be Doing:
Drive post-training research for multimodal decision models and hybrid decision/generative systems addressing research challenges in observability and security.
Own research problems end to end from framing questions through experimentation, model development, and evaluation.
Design and advance post-training methods and training loops for models operating in complex environments.
Build simulated environments, synthetic data generation approaches, and evaluation frameworks for scalable model training and rigorous measurement.
Collaborate with cross-functional teams across Research, Product, and Engineering to translate research advances into scalable Datadog capabilities.
Share research findings with external audiences upon reaching significant project milestones.
Requirements:
PhD in Computer Science, Machine Learning, or related field, or equivalent experience with expertise in AI agents, decision models, post-training, or generative modeling.
Proven track record of driving technically ambitious research at meaningful scale as an individual contributor in industry research labs, startups, academia, or similar.
Extensive hands-on experience designing or implementing training methods, simulated environments, synthetic data pipelines, or evaluation frameworks.
Demonstrated research impact through influential publications, significant model/system contributions, widely used research artifacts, or equivalent achievements.
Experience operating at a senior technical level, leading by influence, mentoring researchers or engineers, and elevating quality of work.
Desire to remain deeply hands-on in research long-term and ability to communicate complex research findings to technical and non-technical audiences.