A Frontend Dev's Guide to Playing with and Learning AI

AI, and more specifically LLMs, have me really pumped. I am more excited today than I was when I first started working with SPAs and feeling the potential that client-side development opened up for us web developers. For me, the excitement comes from being able to leverage very powerful tools to build cool, effective and useful stuff. For the last one and a half years I have buried myself in AI as much as possible, because AI now gives us developers so much more power to build. It is not only challenging trying to get up to speed in this new domain, but even more so trying to stay current with the euphoric level of development and constant changes happening. Build a Foundation Dive deep and keep swimming! It can be intimidating, even for someone who has plenty of development experience, to jump into the realm of [insert whatever aspect of AI that you find challenging here]. Get Started These are some of the resources that helped me get started. https://www.youtube.com/@jamesbriggs - has plenty of "starter" tutorials https://www.youtube.com/@LangChain - a popular library and tooling for building AI apps, they connect very well with developers and make everything super accessible https://www.deeplearning.ai/ - founded by the great Andrew Ng (I have so much gratitude for this man), there are free courses on a wide range of AI/LLM topics https://www.coursera.org/learn/ai-for-everyone - also by Andrew Ng, and I think the title is self-explanatory https://www.coursera.org/specializations/ai-for-good - goes into a more complete lifecycle for building an AI application Keep Building These resources help you go deeper. https://huggingface.co/learn/nlp-course/ - I would agree with the suggestion found on the landing page that you not start with this course, but only after gaining some basic understanding from the courses by DeepLearning.AI https://www.youtube.com/@AndrejKarpathy - if you still do not know who Andrej is, you will soon know and come to appreciate him https://www.youtube.com/@3blue1brown - masterfully produced learning material https://www.dailydoseofds.com/ - a newsletter with tutorials that go a bit deeper, sometimes in multi-part series Staying Current These resources help me stay current with the constant churn of changes. https://www.youtube.com/@WesRoth - for an overview of advancements in the AI space, covers tools, research papers, and general goings-on in the industry https://www.youtube.com/@AZisk - hardware https://www.youtube.com/@NetworkChuck - software and tools, mostly open source stuff you can run locally https://www.latent.space/podcast - the one podcast I make sure not to miss https://newsletter.danielmiessler.com/ - his newsletter covers security and AI, creator of https://github.com/danielmiessler/fabric https://simonwillison.net/ - well known in AI circles, creator of https://github.com/simonw/llm https://bensbites.com/ - a newsletter that shares new tools and community projects Get Your Hands Dirty Install and play with: https://ollama.com/ https://docs.openwebui.com/ https://github.com/danielmiessler/fabric https://github.com/simonw/llm - read the section on "files-to-prompt" https://simonwillison.net/2025/Feb/14/ Starter Project Create an enterprise level RAG application, and I do not mean just copy and paste from any of the number of tutorials online, I mean set something up for hundreds of potential users, for a variety of different documents, documents that are image heavy, running off of a vector DB with your own chunking strategy properly implemented. Use Open WebUI (https://docs.openwebui.com/) as the "UI" and then use a "function" or "pipeline" to integrate your RAG implementation. Consider using Azure Cloud infrastructure, I think you get 3 months free for new accounts, so that you can scale up the availability of GPU intensive computation and avoid having to pay for a GPU-enabled VPS. I learned a lot implementing at my day job what I at first thought was something for amateurs. There are a lot of hurdles that you will need to jump through here, and you will learn a lot in the process.

Mar 24, 2025 - 23:42

A Frontend Dev's Guide to Playing with and Learning AI

AI, and more specifically LLMs, have me really pumped. I am more excited today than I was when I first started working with SPAs and feeling the potential that client-side development opened up for us web developers. For me, the excitement comes from being able to leverage very powerful tools to build cool, effective and useful stuff. For the last one and a half years I have buried myself in AI as much as possible, because AI now gives us developers so much more power to build.

It is not only challenging trying to get up to speed in this new domain, but even more so trying to stay current with the euphoric level of development and constant changes happening.

Build a Foundation

Dive deep and keep swimming! It can be intimidating, even for someone who has plenty of development experience, to jump into the realm of [insert whatever aspect of AI that you find challenging here].

Get Started

These are some of the resources that helped me get started.

https://www.youtube.com/@jamesbriggs - has plenty of "starter" tutorials
https://www.youtube.com/@LangChain - a popular library and tooling for building AI apps, they connect very well with developers and make everything super accessible
https://www.deeplearning.ai/ - founded by the great Andrew Ng (I have so much gratitude for this man), there are free courses on a wide range of AI/LLM topics
https://www.coursera.org/learn/ai-for-everyone - also by Andrew Ng, and I think the title is self-explanatory
https://www.coursera.org/specializations/ai-for-good - goes into a more complete lifecycle for building an AI application

Keep Building

These resources help you go deeper.

https://huggingface.co/learn/nlp-course/ - I would agree with the suggestion found on the landing page that you not start with this course, but only after gaining some basic understanding from the courses by DeepLearning.AI
https://www.youtube.com/@AndrejKarpathy - if you still do not know who Andrej is, you will soon know and come to appreciate him
https://www.youtube.com/@3blue1brown - masterfully produced learning material
https://www.dailydoseofds.com/ - a newsletter with tutorials that go a bit deeper, sometimes in multi-part series

Staying Current

These resources help me stay current with the constant churn of changes.

https://www.youtube.com/@WesRoth - for an overview of advancements in the AI space, covers tools, research papers, and general goings-on in the industry
https://www.youtube.com/@AZisk - hardware
https://www.youtube.com/@NetworkChuck - software and tools, mostly open source stuff you can run locally
https://www.latent.space/podcast - the one podcast I make sure not to miss
https://newsletter.danielmiessler.com/ - his newsletter covers security and AI, creator of https://github.com/danielmiessler/fabric
https://simonwillison.net/ - well known in AI circles, creator of https://github.com/simonw/llm
https://bensbites.com/ - a newsletter that shares new tools and community projects

Get Your Hands Dirty

Install and play with:

https://ollama.com/
https://docs.openwebui.com/
https://github.com/danielmiessler/fabric
https://github.com/simonw/llm - read the section on "files-to-prompt" https://simonwillison.net/2025/Feb/14/

Starter Project

Create an enterprise level RAG application, and I do not mean just copy and paste from any of the number of tutorials online, I mean set something up for hundreds of potential users, for a variety of different documents, documents that are image heavy, running off of a vector DB with your own chunking strategy properly implemented. Use Open WebUI (https://docs.openwebui.com/) as the "UI" and then use a "function" or "pipeline" to integrate your RAG implementation. Consider using Azure Cloud infrastructure, I think you get 3 months free for new accounts, so that you can scale up the availability of GPU intensive computation and avoid having to pay for a GPU-enabled VPS. I learned a lot implementing at my day job what I at first thought was something for amateurs. There are a lot of hurdles that you will need to jump through here, and you will learn a lot in the process.