--- title: Agentic RL: Environments, Trajectories, and the Training... description: Part 12 of a reinforcement learning course focused on training LLM-based agents. Covers what RL environments look like for tool-using LLM agents, trajectories... canonical: https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Agentic RL: Environments, Trajectories, and the Training Loop | daily.dev og:description: Part 12 of a reinforcement learning course focused on training LLM-based agents. Covers what RL environments look like for tool-using LLM agents, trajectories... og:url: https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm og:image: https://api.daily.dev/og/posts/cW82SW5CM.png og:image:alt: Agentic RL: Environments, Trajectories, and the Training Loop og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Agentic RL: Environments, Trajectories, and the Training Loop **[Daily Dose of Data Science \| Avi Chawla \| Substack](https://daily.dev/sources/dailydoseofds)** · 6 min read · 1 upvotes · 0 comments ## Summary Part 12 of a reinforcement learning course focused on training LLM-based agents. Covers what RL environments look like for tool-using LLM agents, trajectories as the unit of training, outcome vs. process-informed scoring, the RULER scoring framework with prefix deduplication, and the credit assignment problem in long-horizon episodes. Includes a hands-on section training a 3B parameter SQL agent using ART and RULER on a free Colab GPU, with no hand-written reward function beyond a one-line correctness bonus. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is RULER used for in agentic reinforcement learning training? RULER acts as a packaged judge for scoring agent rollouts in reinforcement learning, including prefix deduplication to avoid redundant scoring work across trajectories. It is used alongside a tool called ART to train agents, such as a hands-on example training a 3B parameter model to answer database questions using only tool calls on a free Colab GPU with just a one-line correctness bonus as the hand-written reward. _Anyone setting up reward scoring for agent training can track emerging tools like RULER through daily.dev._ ### Why can't prompt engineering alone teach an LLM agent the right tool-use behavior? Prompt engineering only shapes agent behavior indirectly through instructions in the system prompt, relying entirely on the base model's pre-existing capabilities and whatever structure the prompt imposes. It cannot reward desired behavior and penalize unwanted behavior across thousands of rollouts the way reinforcement learning can, which is why RL training loops are introduced for multi-step tool-using agents. _Developers weighing prompting versus RL for agent behavior can follow this tradeoff on daily.dev._ ## Similar posts on daily.dev - [How Top AI Labs Are Building RL Agents in 2026](https://daily.dev/posts/how-top-ai-labs-are-building-rl-agents-in-2026-6czaxnq6j) · Daily Dose of Data Science \| Avi Chawla \| Substack · 1 upvotes · 0 comments - [Reinforcement Learning Environments](https://daily.dev/posts/reinforcement-learning-environments-yi19edoxs) · DigitalOcean Community · 1 upvotes · 0 comments - [How to Fine-Tune LLMs in 2026](https://daily.dev/posts/how-to-fine-tune-llms-in-2026-lx5wel0kf) · Daily Dose of Data Science \| Avi Chawla \| Substack · 1 upvotes · 0 comments - [Mastering Agentic Techniques: AI Agent Reinforcement Learning](https://daily.dev/posts/mastering-agentic-techniques-ai-agent-reinforcement-learning-xsi0iqacc) · NVIDIA Developer · 2 upvotes · 0 comments --- Tags: [#python](https://daily.dev/tags/python), [#ai-agents](https://daily.dev/tags/ai-agents), [#sql](https://daily.dev/tags/sql), [#reinforcement-learning](https://daily.dev/tags/reinforcement-learning), [#agentic-ai](https://daily.dev/tags/agentic-ai) [View this post on daily.dev](https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Agentic RL: Environments, Trajectories, and the Training Loop","url":"https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm"},"datePublished":"2026-07-13T20:26:40.523Z","dateModified":"2026-09-14T06:48:38.053Z","description":"Part 12 of a reinforcement learning course focused on training LLM-based agents. Covers what RL environments look like for tool-using LLM agents, trajectories...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/61bbe7fe3d3ea77e856ada0948b9069c?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/61bbe7fe3d3ea77e856ada0948b9069c?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Daily Dose of Data Science | Avi Chawla | Substack","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Daily Dose of Data Science | Avi Chawla | Substack","logo":"https://media.daily.dev/image/upload/s--4IHQgTOw--/f_auto/v1710503712/logos/dailydoseofds","url":"https://daily.dev/sources/dailydoseofds"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"python,ai-agents,sql,reinforcement-learning,agentic-ai","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Daily Dose of Data Science | Avi Chawla | Substack","item":"https://daily.dev/sources/dailydoseofds"},{"@type":"ListItem","position":3,"name":"Agentic RL: Environments, Trajectories, and the Training Loop"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/agentic-rl-environments-trajectories-and-the-training-loop-cw82sw5cm#faq","mainEntity":[{"@type":"Question","name":"What is RULER used for in agentic reinforcement learning training?","acceptedAnswer":{"@type":"Answer","text":"RULER acts as a packaged judge for scoring agent rollouts in reinforcement learning, including prefix deduplication to avoid redundant scoring work across trajectories. It is used alongside a tool called ART to train agents, such as a hands-on example training a 3B parameter model to answer database questions using only tool calls on a free Colab GPU with just a one-line correctness bonus as the hand-written reward. Anyone setting up reward scoring for agent training can track emerging tools like RULER through daily.dev."}},{"@type":"Question","name":"Why can't prompt engineering alone teach an LLM agent the right tool-use behavior?","acceptedAnswer":{"@type":"Answer","text":"Prompt engineering only shapes agent behavior indirectly through instructions in the system prompt, relying entirely on the base model's pre-existing capabilities and whatever structure the prompt imposes. It cannot reward desired behavior and penalize unwanted behavior across thousands of rollouts the way reinforcement learning can, which is why RL training loops are introduced for multi-step tool-using agents. Developers weighing prompting versus RL for agent behavior can follow this tradeoff on daily.dev."}}]} ```