Senior Applied Scientist, Generative Simulations

Datadog; Paris; On-site; EUR 120,000 - 150,000; Full time

Benefits offered by this job

RSUs/ESPP

Professional development

Mentor program

Inclusive culture

Inclusion Talks

Spring Health

Global benefits

Job summary

Datadog seeks its first applied scientist for GenSim to define methodology for simulated environments and post-training data. You will shape data quality, realism, and evaluation approaches for agent and LLM applications, collaborating with Bits AI SRE and model training teams.

The role emphasizes producing production-grade, scalable environments and guiding how data is used in post-training and evaluation, with strong mathematical and ML foundations.

Qualifications

• PhD, MS or equivalent research experience in a scientific field with strong applied mathematics.
• 6+ years of applied science or ML engineering experience, including setting technical direction.
• Hands-on experience with LLMs and agent post-training data, and how it is created and validated.
• Strong programming in Python and ability to ship production systems.
• Ability to collaborate across engineering and science teams.
• Comfort with ambiguity and making technical calls when paths aren’t defined.

Responsibilities

• Own the applied science direction for GenSim: set methodology and forward-looking calls.
• Define, measure and raise post-training data quality against real distributions.
• Close realism gap; drive research and engineering for realistic, imperfect environments.
• Build scalable, production-grade systems; output is a retrainable synthetic environment.
• Determine best data application in LLM post-training and evaluation.
• Work cross-functionally with Bits AI SRE, model training, and evaluation teams.

Skills

LLM post-training data

Python programming

Production software engineering

Cross-functional collaboration

Ambiguity tolerance

Education

PhD or MS in a scientific field

Job description

Datadog seeks its first applied scientist for GenSim to define methodology for simulated environments and post-training data. You will shape data quality, realism, and evaluation approaches for agent and LLM applications, collaborating with Bits AI SRE and model training teams.

The role emphasizes producing production-grade, scalable environments and guiding how data is used in post-training and evaluation, with strong mathematical and ML foundations.