About the role#
This master's thesis project focuses on System-One models for wireless decision-making. You will explore models that map network states and queries to a calibrated distribution over an action set. This approach aims to replace token-by-token generation with a faster execution engine for agentic cellular modems.
What you'll need#
- Current enrollment in a master's program for Computer Science, Artificial Intelligence, Machine Learning, Software Engineering, or a related field.
- Interest in agentic AI systems.
- Strong English skills, both spoken and written.
- Ability to work well both independently and in a team.
- High attention to detail and a commitment to quality.
- Python programming experience is preferred.
- Knowledge of LLMs, AI agent frameworks, prompt engineering, or reinforcement learning is an advantage.
What you'll do#
- Formalize wireless control tasks as System-One queries.
- Build and test a model calibration pipeline using open wireless datasets or simulated state-action traces.
- Design a confidence-threshold escalation policy and benchmark it against existing baselines.
- Analyze the trade-offs between quality, latency, and calibration across different decision tasks.
Location & details#
- Location: United States.
- Term: Spring 2027.
- Modality: On-site.
- This is a paid, full-time position.


