{"slug": "hybrid-and-local-ai-course-at-deeplearning-ai", "title": "Hybrid and Local AI course at DeepLearning.AI", "summary": "JetBrains and DeepLearning.AI released a free course, 'AI Coding Workflows: Hybrid to Local,' teaching developers how to use hybrid and local AI models in PyCharm, covering techniques like subagents, model routing, and cost reduction. The course demonstrates moving from Claude Code with Anthropic models to OpenCode with OpenRouter and DeepSeek, and finally to fully local setups using LM Studio with Gemma 4 12B and Qwen 3.5 27B, showing that local models can perform well with proper guardrails.", "body_md": "[Learning Courses](/pycharm/category/learning-courses/)\n\n[News](/pycharm/category/news/)\n\n# Hybrid and Local AI course at DeepLearning.AI\n\nOpen weight models are having a moment, driven by control, choice, and cost. Hybrid and local AI are now getting serious looks, so JetBrains teamed up with [DeepLearning.AI](https://www.deeplearning.ai) on a free [AI Coding Workflows: Hybrid to Local](https://learn.deeplearning.ai/courses/ai-coding-workflows-from-cloud-to-local/lesson/napfi1/introduction) course that covers the ideas and options.\n\nThe course is now available and uses [PyCharm](https://www.jetbrains.com/pycharm/) and its AI Chat. Here’s a peek into the course.\n\n## Claude Code: Subagents and cheaper models\n\nWe start the course with, well, not-local. Instead, we use what you already know – Claude Code and its Anthropic models – to introduce some of the techniques and “levers” that help bring choice, control, and even cost reduction. (Yes, I wrote emdashes.)\n\nWe did a previous course on [Spec-Driven Development](https://www.deeplearning.ai/courses/spec-driven-development-with-coding-agents) (SDD) so of course, we wanted to start there. Smaller models struggle with big, open-ended “vibe coding.” Dividing and bounding the work keeps smaller models on track. Important note: this course’s example app is *really* basic. You might say “that’s too easy.” But that’s part of the takeaway: big brain models can do the upfront work, forming right-sized steps for smaller models.\n\nWe then illustrate this division with a Claude Code *subagent*. The main chat prompt implements each roadmap phase in a fresh subagent, to better manage context. This then gives the payoff: a cheaper model for the implementer. Use a “big brain” (Opus) for main conversation thinking and a “little brain” (Haiku) for implementation.\n\nEach lesson finishes with metrics about the change in tokens, turns, cost, and estimated wall time. Which brings us to the main course goal: learning the *ideas* instead of the specifics, which change weekly.\n\n## New agent, inference, and model\n\nThat covers the four levers:\n\n- Specs shaped for the model size\n- Specialist subagents to divide work\n- Cheaper models for the routine work\n- Collect metrics as evidence to guide thinking\n\nThe course then introduces choice and control:\n\n- New agent: OpenCode\n- New inference router: OpenRouter\n- New model and inference host: DeepSeek (via OpenRouter) by moving to a new agent (OpenCode) using inference routing (OpenRouter) to inference hosting and models (DeepSeek)\n\nWe first move to OpenCode, running in PyCharm. JetBrains wants our IDEs to be *open platforms* for agents and models. This makes the move from Claude Code to OpenCode straightforward: it’s the same UI. We add OpenRouter (a paid step), connect it to OpenCode, and choose DeepSeek as a model.\n\nNext we repeat our sequence: all in one chat, then context isolation using a subagent. But this time, with a different agent and model.\n\nWe finish by making a dedicated implementer subagent in Markdown. This gives quite a number of levers of control: in the frontmatter for mandatory controls, and in the subagent body for “persuasion” guidance. Most importantly, we have the implementer use the smaller DeepSeek v4 Flash model as the “little brain.”\n\nCompared to the Claude Code version, the metrics were, unsurprisingly, a lot cheaper.\n\n## Hybrid and Local\n\nNow for the main attraction: for routine development, can we do some – or even all – of the work locally?\n\nWe start with a lesson on setting up local AI: LM Studio as the inference server and Gemma 4 12B as the local model, targeting a 32 GB laptop.\n\nWe then configure the implementer subagent to use this local Gemma 4 model, promoting DeepSeek v4 Flash from last lesson’s “little brain” up to “big brain.” The results? Quite good, as it turns out.\n\nThen the big test: fully local, with Qwen 3.5 27B as the “big brain.” The results: better than expected, showing that guardrails help.\n\nHow did hybrid and local do? Both of these lessons finish with a review of the metrics. That’s one of the big course takeaways: look at the evidence. You can see how small models struggle, and see the effect of helping them succeed.\n\n## Hybrid and Local AI Are Heating Up\n\nMuch thanks to [DeepLearning.AI](http://deeplearning.ai/) both for working with us again and for pushing to get this out fast. This topic is now red-hot in the news: Sovereign AI, privacy and security, and of course cost. The innovations are coming really fast and it is important to have a gentle introduction to the fundamentals.\n\nWe’ll do more updates here on Local AI for control, choice, and cost. Most of all, we at PyCharm believe in the human-in-the-loop. Stay tuned for more on this.\n\n#### Subscribe to PyCharm Blog updates", "url": "https://wpnews.pro/news/hybrid-and-local-ai-course-at-deeplearning-ai", "canonical_source": "https://blog.jetbrains.com/pycharm/2026/08/hybrid-and-local-ai-course-at-deeplearning-ai/", "published_at": "2026-08-13 10:01:32+00:00", "updated_at": "2026-08-13 11:53:17.559085+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-tools", "ai-products", "ai-infrastructure"], "entities": ["JetBrains", "DeepLearning.AI", "PyCharm", "Claude Code", "Anthropic", "OpenCode", "OpenRouter", "DeepSeek"], "alternates": {"html": "https://wpnews.pro/news/hybrid-and-local-ai-course-at-deeplearning-ai", "markdown": "https://wpnews.pro/news/hybrid-and-local-ai-course-at-deeplearning-ai.md", "text": "https://wpnews.pro/news/hybrid-and-local-ai-course-at-deeplearning-ai.txt", "jsonld": "https://wpnews.pro/news/hybrid-and-local-ai-course-at-deeplearning-ai.jsonld"}}