Applied Scientist
Hybrid
Full Time
#Engineering
#Security
#AI
#Machine Learning
#Distributed Systems
#Unity
#Python
#Java
#C++
At Improbable U.S. Defense & National Security, you will help customers use our synthetic environment platform to plan and train for the most complex security challenges. Our team in Arlington, VA, and other U.S. locations applies multiplayer gaming technology to real-world defense and national security problems, creating some of the most realistic virtual worlds available today.
What you'll be doing
- Combine leading models to design, build, and refine modeling and simulation applications that address defense and security needs.
- Prototype and iterate on models of real-world systems so customers can examine past events, interpret current conditions, and anticipate future outcomes.
- Partner with engineers and product teams to implement high-performance models in production environments and strengthen our core technology platform.
What you'll bring
- A degree in a scientific, engineering, or mathematical field, ideally with a computational focus, plus a strong record of delivering technical, data-rich projects for external clients.
- Hands-on experience with statistical modeling and analysis, agent-based modeling, micro-simulation, computational social science, or modeling socio-technical systems.
- Practical coding skills in at least one relevant language, with the flexibility to work across Python, Java, C++, and related tools; familiarity with game engines such as Unity or Unreal is a plus.
- Experience working with complex, real-world data, including geospatial data, and the ability to assess data quality, define modeling assumptions, and communicate results clearly to clients through writing, visualization, and in-person discussion.
- Interest in probabilistic programming and a willingness to stay current with cutting-edge research while rapidly building expertise in new subject areas.
What you'll get
- Hybrid work arrangements that combine time in the office with remote flexibility.





