below marketmost similar roles pay hereabove market
This role pays more than
80%
of similar roles. Most pay
$177,250–$255,000
— the blue band above.
At the midpoint, this role pays about
$274k
versus about
$216k
for comparable roles.
Based on 240 similar postings.
Employer
About Datadog
Datadog, Inc. is an American company that provides an observability service for cloud-scale applications, providing monitoring of servers, databases, tools, and services, through a SaaS-based data analytics platform.
Datadog currently has
73 open roles
on FindRole.
Listed pay typically runs
$192,000–$240,000
across 71 roles with salary data.
Research Manager - Post-Training for Agents
As a Research Manager within Datadog AI Research, you will lead a team focused on fundamental research in post-training and autonomous agents while maintaining a deep involvement in technical direction. You will guide researchers and engineers to advance reinforcement learning, simulated environments, synthetic data generation, and evaluation methods for agents operating across complex real-world systems. Your daily work involves designing training loops, building evaluation frameworks, and collaborating with product and engineering teams to translate research into scalable capabilities. You must possess expertise in reinforcement learning, generative modeling, and post-training methods. The role requires technical leadership to solve problems grounded in observability and security, with a focus on producing research publications and open-source model artifacts for the AI era.
Lead research in agent post-training and reinforcement learning to set technical directions for research programs.
Mentor and develop a team of researchers and research engineers while remaining hands-on in technical problem-solving.
Design and advance reinforcement learning approaches and training loops for agents in complex environments.
Build simulated environments, synthetic data generation approaches, and evaluation frameworks for scalable agent training.
Collaborate with product and engineering teams to translate research advances into scalable Datadog capabilities.
Contribute to research publications and present findings at top-tier conferences like NeurIPS, ICLR, and ICML.
Open-source key model artifacts and benchmarks to the broader research community.
What we're looking for
Hold a PhD in Computer Science, Machine Learning, or a related field, or have equivalent experience.
Possess deep expertise in reinforcement learning, AI agents, post-training, or generative modeling.
Demonstrate the ability to lead technically ambitious research at meaningful scale in an industry, startup, or academic setting.
Have extensive hands-on experience designing or implementing reinforcement learning systems, agent training methods, simulated environments, synthetic data pipelines, or evaluation frameworks.
Maintain a track record of research impact through influential publications, significant model or system contributions, or widely used research artifacts.
Have experience leading, mentoring, and developing teams of researchers or engineers while remaining deeply hands-on in technical direction.
Ability to communicate complex research findings effectively across both technical and non-technical audiences.
Large Language Models
Reinforcement Learning
PyTorch
Multimodal Reasoning
Distributed Training
Post-Training
Prompt Optimization
Agentic Systems
Generative Modeling
Computer Vision
Machine Learning
Context Compaction
Memory Systems
Tool-use
Data Infrastructure
Large Language Models
Reinforcement Learning
PyTorch
Multimodal Reasoning
Distributed Training
Post-Training
Prompt Optimization
Generative Modeling
Agentic Systems
Context Compaction
Memory Systems
Computer Vision
Machine Learning
Data Infrastructure