In today's fast-evolving tech landscape, staying up to date with the latest developments in artificial intelligence (AI) is crucial for software engineers. AI has become integral to numerous industries, from healthcare to finance and beyond. One of the best ways to keep your AI skills sharp is by getting involved in open source projects. In this blog post, we'll explore some open source projects that every software engineer should consider to stay current with AI trends and innovations.

### 1. Llama 2 
It is an open source large language model provided by Microsoft and Meta. If you want to build something like ChatGPT without using OpenAI's API, then llama2 is one of the best options for you. It is free for research and commercial use.

###### Links:
[https://github.com/facebookresearch/llama](https://github.com/facebookresearch/llama)  
[https://github.com/ggerganov/llama.cpp](https://github.com/ggerganov/llama.cpp)  
[LLaMA2 with LangChain by Sam Witteveen](https://www.youtube.com/watch?v=cIRzwSXB4Rc)   
[Llama2 in LangChain by James Briggs](https://www.youtube.com/watch?v=6iHVJyX2e50)  
[How To Install LLaMA 2 Locally with TextGen WebUI by Matthew Berman](https://www.youtube.com/watch?v=k2FHUP0krqg)



### 2. GPT Engineer
"GPT Engineer is made to be easy to adapt, extend, and make your agent learn how you want your code to look. It generates an entire codebase based on a prompt"(AntonOsika, 2023). Ideally if we give proper instruction, it can be used to write a whole software. Personally, I think this will help reduce time spent on typing for a Software Engineer but I think it is not perfect yet.

###### Links:
[https://github.com/AntonOsika/gpt-engineer](https://github.com/AntonOsika/gpt-engineer)  
[How To Install GPT-Engineer by Matthew Berman](https://www.youtube.com/watch?v=ceMuK0xUtSY)  
[GPT Engineer by Dave Ebbelaar](https://www.youtube.com/watch?v=gWy-pJ2ofEM)  



### 3. Auto-GPT
Auto-GPT moves the AI industry one step closer to true artificial general intelligence. Auto-GPT is based on the language model GPT-4 and can be used in the same ways as chat bots like ChatGPT, Bard. But it automates tasks to achieve goals faster. When a user inputs a prompt, Auto-GPT will create sub-tasks for itself to process further until ultimate goal is reached. It also has features like Internet Access and Short Term Memory. Personally, I think this one is also not perfect yet. 

###### Links:
[https://github.com/Significant-Gravitas/Auto-GPT](https://github.com/Significant-Gravitas/Auto-GPT)  
[How To Use AutoGPT by Howfinity](https://www.youtube.com/watch?v=v-5AWQlTFw8)  
[Install Auto-GPT Locally by Matt Wolfe](https://www.youtube.com/watch?v=0m0AbdoFLq4)
 

### 4. MetaGPT
MetaGPT takes a one line requirement as input and outputs user stories / competitive analysis / requirements / data structures / APIs / documents, etc. Internally, MetaGPT includes product managers / architects / project managers / engineers. It provides the entire process of a software company along with carefully orchestrated SOPs.

###### Links:
[https://github.com/geekan/MetaGPT](https://github.com/geekan/MetaGPT)  
[How To Install MetaGPT by Matthew Berman](https://www.youtube.com/watch?v=uT75J_KG_aY)  
[MetaGPT Setup by Prompt Engineering](https://www.youtube.com/watch?v=nqZlTV_L6Ao)  



### 5. Stable Diffusion
Stable Diffusion is a text-to-image diffusion model and it is like the opensource alternative for Midjourney and DallE. The model behind it is relatively lightweight and runs on a GPU with at least 10GB VRAM. Personally, I think it is very good except when you generate faces with mouth open, the teeth are messed up. 

###### Links:
[https://github.com/CompVis/stable-diffusion](https://github.com/CompVis/stable-diffusion)  
[https://github.com/AUTOMATIC1111/stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui)  
[https://github.com/lllyasviel/Fooocus](https://github.com/lllyasviel/Fooocus)  
[Stable Diffusion Crash Course by Freecodecamp.org](https://www.youtube.com/watch?v=dMkiOex_cKU)  
[Advanced Stable Diffusion Features in Fooocus by HolidayEffects](https://www.youtube.com/watch?v=8Rh0TsZeF6c)



### 6. YOLOv8
Even though generative models have been hot news, Artificial Intelligence is not just about Generation. YOLO, which stands for "You Only Look Once," is a popular deep learning object detection model used in computer vision tasks. YOLO models have been widely used in various applications, including autonomous vehicles, surveillance systems, robotics, and more, where real-time object detection and classification are crucial. They offer a good balance between accuracy and speed, making them valuable tools in computer vision research and practical applications. 

###### Links : 
[https://github.com/ultralytics/ultralytics](https://github.com/ultralytics/ultralytics)  
[Train Yolov8 on a custom dataset by Computer vision engineer](https://www.youtube.com/watch?v=m9fH9OWn8YM)  
[Simple YOLOv8 Object Detection with Webcam in Real-time by Nicolai Nielsen](https://www.youtube.com/watch?v=O9Jbdy5xOow)


### 7. LangChain
LangChain is a framework that makes it easier to build applications using language models. It allows applications to do two key things:
Be context-aware: This means connecting a language model to various sources of context, like prompt instructions, a few-shot examples, or relevant content, to help the model provide better responses.
Reason effectively: LangChain helps applications rely on a language model to make informed decisions, such as how to respond based on the given context and what actions to take.

###### Links:
[https://github.com/langchain-ai/langchain](https://github.com/langchain-ai/langchain)  
[https://github.com/msoedov/langcorn](https://github.com/msoedov/langcorn)  
[LangChain Crash Course by codebasics](https://www.youtube.com/watch?v=nAmC7SoVLd8)  
[LangChain Complete Tutorial by Coding Crashcourses](https://www.youtube.com/watch?v=a89vqgK-Qcs)



### 8. Hugging Face
Hugging Face is a company and open-source community that is well-known for its contributions to natural language processing (NLP) and machine learning. They provide a wide range of NLP-related resources, including pre-trained language models, libraries, and tools. Their Transformers library is widely used in the NLP community for working with transformer-based models. 

###### Links:
[https://github.com/huggingface](https://github.com/huggingface)  
[https://huggingface.co/models](https://huggingface.co/models)

### 9. Bark
Bark is a transformer-based text-to-audio model. It can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying. 

###### Links:
[https://github.com/suno-ai/bark](https://github.com/suno-ai/bark)  
[Bark AI Full Tutorial by 1littlecoder](https://www.youtube.com/watch?v=lDeeSjShilQ)  
[Voice Cloning with Bark by Prompt Engineering](https://www.youtube.com/watch?v=o8-1hb7hFTI)  


### 10. LocalGPT
LocalGPT is an open-source initiative that allows you to converse with your documents with complete privacy. With everything running locally, you can be assured that no data ever leaves your computer. It can Seamlessly integrate a variety of open-source models and is built on top of LangChain. 

###### Links:
[https://github.com/PromtEngineer/localGPT](https://github.com/PromtEngineer/localGPT)  
[Video by Prompt Engineering to get started with it](https://www.youtube.com/watch?v=lbFmceo4D5E)  

### 11. h2oGPT
Query and summarize your documents or just chat with local private GPT LLMs using h2oGPT. It can summarize private offline database of any documents such as PDFs, Excel, Word, Images, Code, Text, MarkDown. And supports a variety of models such as LLaMa2, Falcon, WizardLM, LORA etc. Also has UI and CLI too.

###### Links:
[https://github.com/h2oai/h2ogpt](https://github.com/h2oai/h2ogpt)  
[https://evalgpt.ai/](https://evalgpt.ai/)  
[Video by Matthew Berman to get started with it](https://www.youtube.com/watch?v=H8Dx-iUY49s)  


### 12. ChatDev Communicative Agents for Software Development
ChatDev stands as a virtual software company that operates through various intelligent agents holding different roles, including Chief Executive Officer , Chief Product Officer , Chief Technology Officer , programmer , reviewer , tester , art designer . These agents form a multi-agent organizational structure and are united by a mission to "revolutionize the digital world through programming." The agents within ChatDev collaborate by participating in specialized functional seminars, including tasks such as designing, coding, testing, and documenting. The primary objective of ChatDev is to offer an easy-to-use, highly customizable and extendable framework, which is based on large language models (LLMs) and serves as an ideal scenario for studying collective intelligence. 

###### Links:
[https://github.com/OpenBMB/ChatDev](https://github.com/OpenBMB/ChatDev)   
[Video By Mathew Berman on How To Get Started With It](https://www.youtube.com/watch?v=yoAWsIfEzCw)  


In today's fast-evolving tech landscape, staying up to date with the latest developments in artificial intelligence (AI) is crucial for software engineers.

18296.jpg

12 Must-know Open-Source Projects on Artificial Intelligence (As a Competitive Software Engineer)

I’ve been building AI agents non-stop since November 2024, and I’d like to share some of the things I’ve learned and how I approach building them.

I’m the kind of person who prefers doing things from scratch, especially when working with technologies that are new to me. That means I don’t rely on popular libraries—other than the standard SDKs from LLM providers. So, if you’re expecting fancy frameworks, there’s nothing to see here. This is all about pure logic.

My agent architecture : 
IO <-> Agent Router <-> Business-Specific Agent <-> LLM Service <-> Providers

IO: 
This is the standard input/output layer, handling request validation, authentication, etc.

Agent Router:   

Depending on the business logic, I either route agents by project/task status or, in some cases, use tool calls from a primary agent (e.g., a planner agent) to handle routing.

Business-Specific Agent:  

This is where the agent’s core logic lives. It iterates (loops) until the task is completed or until it needs to send a message back to the user. In this layer, I define which model strategies to use, what prompts to provide, which tools to provide and the context to maintain across iterations. To avoid runaway loops, I also implement a safety max iteration limit—if the agent reaches this limit, it pauses and prompts the user for approval before continuing.

Context Handling:  

For agents that primarily work with tools, I usually provide the 2-3 most recent messages in raw form and prior messages in summarized form to keep context efficient. Agents also have access to a context set/get tool that allows them to save multi-step work and retrieve it later when needed. This is crucial for handling workflows that span multiple steps or require persistent memory during a task.  

LLM Service:  

Acts as a router for different LLM providers. Here, I handle API request logging, error logging, token usage tracking, and retries. Based on error codes, I might retry with the same model strategy or switch to a different one. Some agents support multiple models, so retrying with alternate strategies is common.

Providers:  

I primarily use Google Gemini and OpenAI models because they’re free during development (via Gemini’s free tier and GitHub Models API). Claude models are undeniable when building coding agents but other than that, I use Google and OpenAI models.

Mock Provider:  

I also use a mock provider that returns predefined responses based on the user’s message count or specific keywords. This is extremely helpful for developing workflow agents—you get predictable, fast responses and can focus on the application side without waiting for actual LLM outputs. Highly recommended.

Human Provider:   

There’s also a “manual provider” where I send responses manually through a special route, effectively acting in place of the AI agent. This is useful for reviewing tool call results and the context provided to the LLM. Building AI agents is all about context and tools—if the tools you give your agents aren’t effective, they’ll fail no matter which model you use. This one’s also highly recommended.

Evals:  
I treat evals like test cases in traditional software engineering. I group eval sets into levels: easy, difficult, edge cases, etc. Each eval defines the task for the agent, success criteria, expected scores, and the minimum/maximum tool calls. I don’t do anything fancy here—I run them manually to see if the agents behave as expected. (There might be better processes I haven’t explored yet.)

This is the way I build AI agents today. 

Thank You.




I’ve been building AI agents non-stop since November 2024, and I’d like to share some of the things I’ve learned and how I approach building them...

Artificial Intelligence

Browse All Posts

12 Must-know Open-Source Projects on Artificial Intelligence (As a Competitive Software Engineer)

Here’s how I build AI agents in 2025