MAESTRO: An AI coach that shapes practice and rewards for smarter traffic control

Kari Jaaskelainen

25 Nov 2025 — 1 min read

Meet MAESTRO: an AI coach for teamwork

Training teams of AI agents is hard: you must design the right "score" (rewards) and the right practice plan (curriculum). Get either wrong and learning stalls. MAESTRO uses a large language model (LLM) as an offline training architect—so it guides learning without slowing real-time decisions.

Curriculum generator: creates diverse, performance-driven traffic scenarios that steadily raise the challenge.
Reward synthesizer: auto-writes executable Python reward functions tailored to each stage.

Because the LLM is used only during training, deployment stays fast and cheap. MAESTRO plugs into a standard multi-agent RL backbone without changing inference cost.

In a 16-intersection Hangzhou traffic simulation, MAESTRO improved average return by +4.0% (163.26 vs. 156.93) and boosted stability by 2.2% (Sharpe 1.53 vs. 0.70) over a strong curriculum baseline. Ablations show the combo of LLM-made curricula + rewards works best.

Big picture: LLMs can be great designers for cooperative AI—setting the practice and the scoring—while the agents handle the real-time control.

Paper: https://arxiv.org/abs/2511.19253

Paper: https://arxiv.org/abs/2511.19253v1

Register: https://www.AiFeta.com

AI ReinforcementLearning MARL MultiAgent LLM MachineLearning SmartCities Traffic

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added

Meet MAESTRO: an AI coach for teamwork

Read more

Meet GRACE: a moral governor for safer, more transparent AI

Why some fine-tuned LLMs miss phishing—and how to fix it

AI that explains itself—by following the science

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A