Open-source LLMs: Uncensored & secure AI locally with RAG (Udemy.com)

Just Amazing!!!! Arnie Really goes into details with ups and down of every path and doesn't skip over any essential part . This course is basic enough to be accessible by literally anyone !!! Most amazing thing is we learn a lot in very amount of time . One can easily drag these types of courses to 40 hours and waste time but Arnie didn't .

This is a good course, covering a wide range of topics including Sensored and Unsensored LLMs, Local Deployment, AI Agents, and Fine-Tuning. What I particularly appreciate about this course is the instructor’s insights beyond the core topics. It doesn’t just focus on tools and platforms but also dives into the details that learners might overlook during their journey. This course serves as a strong starting point for anyone looking to explore LLMs. It effectively covers the essentials and helps learners understand the complexities relevant to the field.

I took this course to get a practical overview of the open-source LLM ecosystem and how to run and use models locally, and in that regard it delivers very well. It covers a wide range of topics and tools — LM Studio, Ollama, Anything LLM, RAG, vector databases, function calling, basic agents, fine-tuning, and even hardware considerations — and does so in a clear, step-by-step, very approachable way. It’s especially useful if you want a “big picture” plus working setups rather than just theory. That said, this is more of a broad, tool-oriented course than a deep programming course, and some parts depend on fast-moving tools that can become outdated. I would have liked a bit more depth in some advanced areas (RAG and agents in particular), but overall it’s a solid course for developers or technical users who want to understand and start using open-source LLMs locally without relying only on closed-source APIs.

It started out good but I felt like as the course progressed the content became more rushed and less substantive. The instructor is a good presenter, but is sometimes a little bit repetitive. And asking at the end of every section to share the course was a little bit annoying. With German as his first language, Arnie has very understandable English. There were a few mistakes, but nothing that detracts from understanding or distracts from learning. He presents well and his recordings, screen captures, and so on were very professionally done and easy to follow. I wanted more in-depth content, but I may have misunderstood the purpose of the course. If you are interested in setting up a functional LLM server in your house, you will learn enough to be able to do that in this course, but I would not say that the way things are installed is a professional installation (there is no docker, no nvidia toolkit, no linux, etc.). The course takes place in Windows land, and most applications run in the foreground. For a course on local AI, there was an unexpected emphasis on API models and API services. For local AI, I do not care about these. However, the instructor presented them mostly in compare & contrast type discussions, so it wasn't useless; just not really what I was looking for. I think the course content could probably have been covered in less time. There are some videos that felt a bit rushed. I was also hoping he would cover LoRA training, which is more useful than fine-tuning. All that being said, I did learn some things in the course. I did set up a local LLM server and with the content covered and previous devops experience, I had no trouble setting things up the way I wanted (linux server, docker containers, etc.). A really nice addition to this course would be to show how to integrate a local LLM system with development tools via API keys. For example, VSCode, Jetbrains, etc., with continue.dev plugin, can connect to local LLMs for code completion and agentic code editing. It would also make sense to talk about MCP servers and integrating custom code into local LLMs. This course was VERY surface level in terms of what is possible. And there are no projects or interesting use cases explored. There are a lot of demonstrations of things working, but not much creative exploration or discussion on how it works under the hood. It is a beginner-level course. Also, I had a real, genuine laugh when he demonstrated Moshi chat. His conversation with the voice bot was unexpected and hilarious. Throughout the course, you'll find that Arnie has a very dry and subtle sense of humor, which I personally appreciate.

Clear, Practical, and Worth Every Minute Arnie’s course on open-source LLMs stands out in all the right ways. He doesn’t just throw concepts at you, he breaks them down step-by-step, shows you exactly how to get things running locally, and keeps everything practical and grounded. What I really appreciated was the balance between theory and hands-on application, from running quantized models on limited hardware to building basic agent flows and understanding prompt engineering. Arnie also made sure to touch on the bigger picture: privacy, licensing, and the real-world tradeoffs between open- and closed-source LLMs. If you're curious about open-source LLMs but don’t know where to start (or how deep you can go), this course is a solid guide. Highly recommended, and looking forward to his next one.

this course begins great in fact I learned a lot from the good parts that I decided to raise the score to 4/5 stars you will indeed learn to use LLMs about prompts with examples, introduced to LM studio,anythingLLM,ollama and flowise but flowise is were this course falls apart. in the flowise section it was not up to date- I would expect a course that it's whole point is to run stuff locally (which was awsome!) to provide a way to create agents with ollama that still works or with any other local llm server, the finetuning and security sections were done in the most lazy way if I hadn't felt that the first part did gave me a lot i would give it a 3 or below the last parts were just a waste of time, explained in the most boring fashion. but the first parts as I said were worth the whole course, understanding how to prompt engineering made my LLM interations way more efficient and experimenting with rag and agents were really good (even though i had to use api key for the multi agents part instead of trying it fully locally) security and jailbreaking should be more extensive in my opinion

My interest in understanding RAG is what drew me to this course. But i got more than that. This is an amazing course, covering a broad swath of topics from open source LLMs, building applications with them using LLM studio, understanding and building RAG applications. The cherry on the top was understanding and building agentic AI based applications using lang chain and Flowise. The course even discusses the vulnerabilities in LLM application like jailbreaking and data poisoning and aspects of data privacy and copy rights. Arnie is a great teacher and like his style of reinforcing the learnings during the wrap up of each session. Bravo Arnie.. you are doing a great service to the community!

Anyone interested in LLMs should take this course. This is an extremely well-put-together course. Arnio is well-versed in all the topics covered here. He explains them in a way so everyone can understand them. He shows good examples and gives clear explanations of what he is doing at each step when putting agents together. Finally, the topic of security comes up throughout the course, and specific sections on just those topics once again very informative. As a developer this is important to me because it showed exactly how some of this could be done and what to look out for when creating agents from various LLMs. This is a must-have resource when working and developing your own AI projects.

Yes it explain a lot from the AI but in the video the trainer repeat over and over the same knowledge. There is a lot AI LLM Models it will be nice it would be more examples RAC application realted to health like diabetese, garming watch csv data from the diabetese sensor ,and good llm to anallyse this data. That can really help sick people. Flowise version chnaged and there is no ChatOllama Function so section 7 58 and 59 need a big update. I mean creator should prepare document with hte versions of the open soource application that we have the same working environment at least for example : nodje version, npm version, flowise version.

The primary goal of this course is to deliver immediate, tangible results. While it may not be obvious at first, the selection of user-friendly, easy-to-learn tools significantly simplifies decision-making and makes the learning process much smoother for beginners in this expansive field. It’s impressive that so much can be accomplished in under 10 hours, and equally remarkable to discover tools like LM Studio and AnythingLLM that enable you to become immediately operational. These tools make complex topics like RAG (Retrieval-Augmented Generation) far more approachable, eliminating the need to build everything from scratch or spend countless hours studying frameworks that may not be as focused on delivering quick results. This emphasis on practicality is, without a doubt, the course's greatest strength. That said, the content can sometimes be a bit more concise than I would have preferred, and I would have appreciated more coverage of both audio and video. Additionally, it would have been more convenient if the resources were listed under the specific lectures they pertain to, rather than in a single document that consolidates all course resources. While I would have liked more depth on RAG, I acknowledge that what is taught is sufficient for achieving practical outcomes. Overall, this is one of the few Udemy courses that I would truly regret missing.