Caltech is a world-renowned science and engineering institute that marshals some of the world''s brightest minds and most innovative tools to address fundamental scientific questions. We thrive on finding and cultivating talented people who are passionate about what they do. Join us and be a part of the diverse Caltech community.
Job Summary
We are seeking a highly motivated Research Scientist with a strong background in networks and distributed systems to work with the Caltech High Energy Physics group and partners at Oak Ridge National Lab in the SciNet project which is part of the Genesis Mission. The position will focus on the development of a science workflow-aware network system supporting data intensive science programs including the Large Hadron Collider program. The position will involve research on AI agents for intelligent, autonomous, and self-optimizing networks for supercomputing and scientific infrastructure.
The researcher will work at the intersection of artificial intelligence, computer networking, high-performance computing (HPC), distributed systems, and autonomous systems. The position will contribute to the development and evaluation of AI-agent architectures capable of monitoring, reasoning about, optimizing, and autonomously managing complex scientific networks.
Essential Job Duties and Key Responsibilities
AI-Agent Research
- Investigate architectures for AI agents operating in networked systems.
- Develop and evaluate LLM-based and/or reinforcement-learning-based network agents.
- Explore multi-agent approaches for network monitoring, optimization, troubleshooting, and orchestration.
- Investigate agent planning, reasoning, tool use, memory, and coordination for network operations.
- Develop methods for safe and reliable autonomous network decision-making.
Network and HPC Research
- Analyze network telemetry, traffic patterns, performance metrics, and system events.
- Study network optimization problems including routing, congestion control, traffic engineering, and resource allocation.
- Investigate interactions between compute, storage, and network resources in HPC environments.
- Evaluate AI-agent approaches using HPC and scientific networking workloads.
Experimental Research and Testbeds
- Develop experimental environments using network simulators, emulators, digital twins, and/or HPC testbeds.
- Implement prototype AI-agent systems and integrate them with networking tools.
- Conduct controlled experiments to compare AI-agent approaches against conventional network-management techniques.
- Collect, analyze, and visualize experimental results.
Software Development
- Develop research prototypes primarily in Python and, where appropriate, C/C++ or other systems languages.
- Work with machine-learning frameworks and APIs.
- Develop interfaces between AI agents and network monitoring/control systems.
- Build reproducible experiment pipelines and research software.
- Use Git/GitHub or equivalent version-control systems and follow good software-engineering practices.
Research Communication
- Contribute to conference and workshop submissions.
- Present research results to interdisciplinary audiences.
- Participate in research meetings with university, national laboratory, and industry collaborators.
Basic Qualifications Required
- PhD in Computer Science, Electrical/Computer Engineering, Data Science, AI/ML, or a related field.
- Strong programming skills, preferably in Python.
- Understanding of computer networking and distributed systems.
- Familiarity with machine learning or artificial intelligence.
- Ability to conduct experiments, analyze data, and communicate technical results.
- Strong written and verbal communication skills.
- Demonstrated ability to work independently and as part of a multidisciplinary research team.
Preferred Qualifications
Experience with one or more of the following is highly desirable:
- Large language models and generative AI
- AI agents or agentic AI systems
- Reinforcement learning
- Multi-agent systems
- Network simulation/emulation
- High-performance computing
- Software-defined networking
- Network telemetry and observability
- Digital twins
- Kubernetes or cloud infrastructure
- PyTorch, TensorFlow, or similar ML frameworks
- Distributed systems
- Network optimization
Required Documents