Outreachy 2026 – Develop a SLM/LLM using RamaLama RAG: Introduce yourself!

[jflory7 (Justin Wheeler)]

Hello everyone, and welcome to the Fedora Project!

Whether this is your first time interacting with an Open Source community or you are a seasoned contributor, we are thrilled you are here. Fedora is a global, diverse community of people working together to build a free and open source operating system, and we are incredibly excited to see your interest in the RamaLama RAG SLM/LLM project!

I am Justin Wheeler, the Fedora Community Architect. Alongside Fernando Fernandez Mancera (@ffmancera), I am one of your two Fedora Co-Coordinators for this Outreachy cohort. I am also excited to be one of your project mentors in this specific cohort, alongside the wonderful Dominik Kawka (@dominikkawka), Carol Chen (@cybette), and Gordon Messmer (@gordonmessmer).

Because we have folks joining us from vastly different time zones and backgrounds, communication is the key to our success. In Fedora, we rely heavily on asynchronous, text-based communication. Before you dive in, we highly encourage you to read through our project guide, Communication in Fedora.

A quick note on how we evaluate contributions: We value quality over quantity. Please know that a high volume of forum posts does not equal a stronger application. Mentors evaluate the quality of your work on your specific project tickets (in Forgejo or GitLab). Please avoid posting in unrelated Fedora topics just to increase your activity metrics—it creates “noise” that makes it harder for mentors to review your work!

Reply template

To help keep our forums organized, please reply directly to this thread to introduce yourself. In your reply, please tell us:

1. Your Open Source journey: What is your current level of experience with open source software? (It is 100% okay if the answer is “none”!)
2. Your “Why”: Why did you choose to participate in Fedora, and what drew you to this specific RamaLama/AI project?
3. Your goals: What is the number one thing you hope to learn or take away from these next few weeks?
4. Just for fun: If you could instantly become a world-class expert in any one non-technical skill or hobby, what would it be and why?

Take a deep breath, don’t be afraid to ask questions on your tickets, and have fun. We can’t wait to meet you!

[chinnisree (Chinni Sree Addagalla)]

Hi everyone! I’m Chinni Sree Addagalla, an Outreachy 2026 applicant for the “Develop a SLM/LLM using RamaLama RAG based off Fedora RPM Packaging Guidelines” project. Excited to be here!

1. Your Open Source Journey:

I’m fairly new to open source contributions. While I have experience building software projects, this Outreachy internship is my first structured open source contribution.

2. Your “Why”:

I chose this project because it genuinely excites me. I’ve always been interested in AI and LLMs, and during my recent externship at Wayfair, I got hands-on experience building multi-step AI agents using Google Gemini and LangGraph with RAG pipelines. When I saw this project was about building a SLM/LLM using RamaLama with RAG, it felt like a perfect fit. I want to take what I’ve learned and apply it in a real open source setting for the first time and hopefully contribute something meaningful to the community.

3. Your Goals:

My number one goal for these next few weeks is to understand how RAG systems can be built and applied in a real-world open source project like RamaLama and to make quality contributions that are actually useful to the community.

4. Just for Fun:

If I could instantly become a world-class expert in any non-technical skill, it would be badminton. I enjoy playing it and love how it keeps you sharp both physically and mentally. There’s something about the fast rallies and quick decision-making on the court that I find really satisfying.

Looking forward to learning and contributing alongside everyone here!

[yrevash (YASH TIWARI)]

Hi everyone! I’m Yash Tiwari, an Outreachy 2026 applicant for the “Develop a SLM/LLM using RamaLama RAG based off Fedora RPM Packaging Guidelines” project. Really glad to be here!

Open Source Journey:
I’ve been contributing to open source for about a year now. A big chunk of that has been in C++ I worked on porting a gestures library to the Avendish runtime during, building media processors for artists in live performance systems, along with that i have contributing to hyprnote rust based on system notes taker for online meetings. This experience taught me a lot.

My “Why”:
AI is where most of my professional work has been over the past year. I’ve built an LLM-powered mobile testing agent using Ollama end-to-end, designed RAG pipelines serving 350,000+ users on a social platform, and worked with LangGraph, Gemini, Qwen, and others across different projects. When I saw this project, it felt less like “learning something new” and more like a chance to go deeper to really stress-test different RAG architectures, agentic search patterns, and graph-based retrieval in a serious open source setting. That’s what pulled me in.

Goals:
I want to go beyond just getting something working. My goal is to understand the fundamentals deeply enough to compare different RAG architectures, experiment with agentic search and graph-based approaches, and contribute something that actually holds up under real usage.

Just for Fun:
I love building robots in my free time, which is probably why I ended up in mechatronics before drifting fully into AI.

Looking forward to contributing alongside everyone here!

[andungu (Anne Ndungu)]

Hi Everyone, my name is Anne Ndungu (@andungu:fedora.im). Super excited to be here collaborate with all of the people from Fedora as well as my fellow Outreachy applicants. Good luck to all of us lets give it our best shot!

Open source Journey:
I dont have much experience with open source but having joined Fedora recently Im finding it great so far. Interacting with everyone in the rooms has been really fun and educational. My focus is on the intersection of Data Science and Open Source. I am currently interested in the RamaLama AI project and I’m comfortable with Git workflows and community engagement via Matrix and Forgejo.

Why
As a Data Scientist, I am drawn to RamaLama because it makes AI models easier to run by making local LLM/SLM usage straightforward through OCI containers. As a person who lives on a farm connectivity sometimes becomes iffy and ramalama solves that making powerful AI models accessible and runnable on local hardware without usage of heavy cloud infrastructure. This project is useful for me and my work and I want to get involved with that!

Goals:
My goal is to master the Packaging side of AI and move beyond just training models to understanding how to containerize, distribute, and maintain AI tools. I want to learn how to ensure that technical documentation is as high-quality and ethical as the code it supports.

Just for Fun:
I love animals and live in an environment with lots of different species of birds. I would love to be able to rehabilitate birds and nurse them back to health! I have already rehabilitated one injured grey headed kingfisher bird and that experience was just great. Especially when the bird grew strong enough to fly away again. Very rewarding experience for me, highly recommend!

Super excited to keep working with all of you!

[iamnjeri (Mercy Njeri Kimaru)]

Hi everyone! I’m Njeri Kimaru, an Outreachy applicant for the Fedora project: Developing an SLM/LLM using Ramalama RAG based on Fedora RPM guidelines. I’m really excited to be here

Open Source Journey
I’m new to open source, and I’ve truly been enjoying the Fedora community so far. Contributing and learning in this space has been an exciting and rewarding experience for me.

Your “Why”
When I came across the Fedora Ramalama/AI project, I immediately felt it was the right fit for me. I’ve been working with RAG systems for some time, and the idea of building and hosting them locally using Ramalama really caught my interest. I’m excited about the opportunity to grow and contribute to this project.

Your Goals
Through this project, I hope to deepen my understanding of Fedora and Ramalama, while continuing to build my skills in AI and machine learning.

Just for Fun
It could definitely be making memes and funny jokes I love memes and the idea that they make people come together.

[theprogram (MatH)]

@iamnjeri

I’ve been working with RAG systems for some time

Can you tell me, what is a RAG system? I’m familiar with LLMs and SLMs, but what is different or special about RAGs?

And Welcome to Fedora!! I used to work with memes, they can really cut through the noise when done properly!

[montana-d (AWA AGHANGU)]

Hi everyone,

My name is Awa Destiny Aghangu, joining from Cameroon as an Outreachy applicant.

My Open Source Journey
No contributions yet, but that’s what I’m here for. Fedora being an established organization with its own standards and culture was actually a big part of why I chose it. As a first step into open source, I’d rather learn within something structured than start somewhere with no guardrails.

My Why
I’ve been using Ollama to run models locally for a while now, so RamaLama felt familiar the moment I came across it. When I read what this project is actually doing, using RAG to check RPM packages against Fedora’s own guidelines, I thought that was a really smart use of AI. It’s not just a demo, it’s solving a real problem for a lot of people. That’s the kind of thing I want to be close to and understand properly.

My Goals
Honestly, I just want things to start clicking. How RAG actually works end to end, how the pieces connect, how a community like this ships something together. I want to go from “I’ve heard of this” to “I understand what’s happening here.”

Just for Fun
Storytelling and public speaking. I’ve always admired people who can make a room care about something. Technical ability opens doors, but being able to communicate well keeps them open.

Looking forward to getting started.

[yrevash (YASH TIWARI)]

Hey MatH,
So LLMs know only what they were trained on their knowledge is frozen at a cutoff. RAG (Retrieval-Augmented Generation) fixes that by giving the model a way to fetch relevant documents at query time and use them as context before generating a response. So instead of relying purely on what it “memorized”, it can look things up on the fly.

Think of it like the difference between answering from memory vs. being allowed to Google something first same model, much better and more current answers.

[cybette (Carol Chen)]

It’s so great to hear from all of you! Thanks for the wonderful introductions.

I’m one of the co-mentors for this Outreachy project. While I’m new to Outreachy, my open source journey has been pretty long. My first encounter with Linux was in 2000 when a university classmate casually mentioned that he spent hours downloading an OS over dial-up and he was eager to try it out. I was intrigued because (1) it wasn’t the virus-prone Windows, (2) it was similar to Unix that we were learnt in college, and (3) you can take the source code and modify it and recompile and distribute it

While in grad school, I pursued machine learning, and built expert systems in Prolog and Lisp. Sometimes I wonder how things would’ve turned out if I stayed in academia and did AI research but I joined Nokia and had the opportunity to contribute to the open source Helix multimedia framework, which was rather pivotal. I started exploring and getting involved in more open source projects, which led me to Red Hat where I’ve been for the past 10 years.

My contributions to FOSS communities have ranged from code contributions to managing the communities, documentation to event planning, mentoring to maintaining websites, and everything in between. Still, everyday I’m learning. I’m sure I’ll be learning as much from y’all as you will (hopefully) learn from me. Let’s fine-tune our knowledge and experiences together! (In this context, RAG is not enough )

Fun fact: I play timpani / percussion in an amateur orchestra. In an alternate reality I would love to be conducting a professional orchestra, bringing together different instruments towards a unified sound. Much like a community of diverse talents coming together, becoming more than sum of its parts to make an open source project like Fedora successful

[montana-d (AWA AGHANGU)]

Truly amazing! Your experience is remarkable!

[kehindeolukosi]

Hi everyone! I’m Kehinde Habeebat Olukosi, an Outreachy 2026 applicant for the “Develop a SLM/LLM using RamaLama RAG based off Fedora RPM Packaging Guidelines” project. Excited to be here!

Your Open Source Journey:
I have contributed to opensource wikimedia projects in the past. I’m eager to contribute to the fedora project as well.

Your “Why”:
I was drawn to Fedora because of its commitment to innovation and community-led standards. The RamaLama project specifically caught my eye because it aligns perfectly with my passion for “frugal innovation.” I recently won a DeepTech hackathon by building a USSD-based AI diagnostic tool for low-resource environments. Seeing RamaLama’s focus on making LLM/SLMs easier to deploy and manage via RAG pipelines feels like the perfect place to apply my Python and Data Science skills to a project that simplifies AI for everyone, while also improving on them.

Your Goals:
My number one goal is to master the integration of RAG systems within the Fedora RPM packaging guidelines. I want to transition from building “standalone” AI projects to understanding how to contribute high-quality, maintainable code and documentation that follows strict open-source community standards.

Just for Fun:
If I could instantly become a world-class expert in any non-technical skill, it would be American Sign Language (ASL). I already have a basic grasp of it, but reaching a world-class level would allow me to bridge communication gaps perfectly. I find the language fascinating!

Looking forward to learning and contributing alongside everyone here!

[theprogram (MatH)]

If anyone is in Nairobi, a previous Outreachy person helps with

Nairobi GNU/Linux Users Group | A lively community of GNU/Linux enthusiasts

They are a great team, just getting going again!

[dohoudanelfavour (Daniel Dohou)]

Hello everyone,

My name is Daniel Dohou, and I am really excited to be here as part of the Fedora Outreachy community this cohort.

My Open Source Journey

My open source journey is still growing, but it has already taught me a lot. I have spent time building, learning, and contributing around software engineering projects, and I am very interested in communities where people work in the open, share knowledge, and improve things together. I would say I am still learning how open source communities work in practice, but I am very comfortable learning publicly, asking questions, and staying consistent until I understand what I need to do.

Why I Chose Fedora and This Project

I chose Fedora because it feels like a community that really values both innovation and collaboration. What drew me most to Fedora is how welcoming and structured the community feels, and I also love that Fedora is open to modern ideas and tools.

I was especially drawn to the RamaLama RAG SLM/LLM project (I kind of like the name “Rama - Lama ”) because it sits right at the intersection of AI, containers, and practical software development. That is a space I find genuinely exciting. I love the idea of working on something that helps simplify how people run and use AI models locally. It feels useful, future-facing, and very aligned with the kind of work I want to keep doing.

My Main Goal

The number one thing I hope to take away from these next few weeks is a better understanding of how to contribute meaningfully in a large open source community while still maintaining quality, clarity, and good communication. I want to become more confident in navigating the workflow, collaborating with mentors, and turning my interest into real value for the project.

Just for Fun

If I could instantly become world-class at one non-technical skill, it would be public speaking or mentorship.

I think it would help me communicate ideas better, share my work more clearly, and become more effective when talking to people in community spaces, team settings, or even when presenting projects. It is one of those skills that would make everything else I do stronger.

I am looking forward to learning from everyone here and contributing as much as I can. Thank you for the warm welcome.

[veronica1w (Veronica Waiganjo)]

Hi everyone!
I’m Veronica Nyambura, an Outreachy 2026 applicant for the “Develop a SLM/LLM using RamaLama RAG based off Fedora RPM Packaging Guidelines"
Happy to be here!

My Open Source journey:
I have been a part of the open-source community, in a project with the Centre for Digital Learning. My experience there has mostly been with the development of Text-to-Speech (TTS) technologies. It’s been a great journey so far, and I’m excited to bring that energy here to Fedora and focus on this project.

My “Why”:
My “why” is honestly just a love for learning. In tech, it feels like there is infinite knowledge and everything moves so fast—there is always something new being built. I want to take every opportunity I can to grow, especially in the LLM space. I’ve worked with models like Llama and Whisper before, and while I’ve heard a lot about RAG systems, I’ve never built one myself. I chose this project because I want to move past just hearing about RAG and actually get my hands dirty building it with RamaLama.

My goals:
My number one goal is to understand how to build and evaluate RAG systems within a production-level open-source ecosystem. I specifically want to learn how to use tools like docling and ramalama to transform static documentation into a dynamic, helpful resource for other contributors.

Just for fun:
If I could instantly become a world-class expert in a non-technical skill, it would definitely be backpacking. I’ve always been so intrigued by people who can just grab a bag and travel the world for multiple months!. I haven’t done it yet, but have you ever looked at something and just knew it was going to make you happy? That’s exactly how I feel about it.

[iamnjeri (Mercy Njeri Kimaru)]

yeah very true @theprogram memes are the way to go.
So RAGs in full is Retrieval Augmented Generation. So you feed your AI model specific documents where it will derive information from, such that it answers you from the documents you feed it but it uses LLM/SLM to derive the information.
For example you can build a RAG system for school students where you feed the model with all the notes and documents they have in school such that when they ask questions they will get their answers from the documents provided with the help of an LLM/SLM.
That’s my understanding of RAGs.

A question to everyone;
Do RAGs strictly answers from the provided documents or does it also answers from it own knowledge by the help of the LLM/SLM?

[iamnjeri (Mercy Njeri Kimaru)]

Thank you so much for this

[iamnjeri (Mercy Njeri Kimaru)]

Wow this is really inspiring​.
In 2000 some of us weren’t even born​.
It’ll be thrilling to learn from you​.

[shehrbanoali6 (Shehrbano Ali)]

Hi everybody!
I’m Shehrbano Ali, a 20 years old ML developer. I am thrilled to be an applicant for this Feroda Outreachy 2026 cohort.

My Open Source, AI Journey:

I am a self taught Machine Learning model developer. Since completing my FSc Pre-Engineering in 2023, I’ve been on a deep dive into Python and AI.
Recently, I expanded my knowledge into AI automation, LLMs, and prompt engineering. This expertise allowed me to develop an AI Receptionist; specifically, an AI customer support voice agent that revolves entirely around LLM and RAG structures. My work is specifically focused on using free-tier, local components to create accessible tools that help people. Moving from these solo projects to a massive open source community like Fedora gives me immense hope.

Why I Chose This Project:

I chose this project as I believe technology should solve real world problems. Whether it’s a real-time agent helping a customer or an AI assistant helping a developer navigate a massive, 5,000 page rulebook I’m eager to apply my RAG experience to help the Fedora community move past the theory and build high impact systems using RamaLama.

My Goal:

Having already built solo projects using RAG, my primary goal now is to scale that knowledge. I want to master tools like docling and ramalama to transform static, complex technical documentation into a dynamic, helpful resource for contributors worldwide.

Just for Fun:

If I could instantly become a world-class expert in a non-technical skill, I would be a Politician; the rare, legendary kind who actually stands for the people! I’ve always felt a strong pull to stand up against unfairness and protect those who need it most. My version of politics wouldn’t be about power; it would be about the right to speak up without, you know, the fear of being ‘deleted’ from the server. I basically want to ‘debug’ the system and deploy a permanent patch for social justice. If I can figure out a complex RAG pipeline, surely I can handle a parliament… right?

[utkarsh-mishra]

Hi everyone, I’m Utkarsh Mishra, an Outreachy applicant for the “Develop a SSL/LLM using Ramalama RAG based off Fedora RPM Packaging Guidelines” project. Really excited to be in such a community.

 1. My Open Source Journey :

Being real I’m new to open source. I have always chased the concept of how things work internally so I studied java and learned some backend concepts and currently I’m pursuing a Data Science and Applications program from IIT Madras. I have not much history to tell about my open source but I’m working on building one. Through Outreachy I started exploring open source looking forward to contribute into fedora and many such projects.

 2. My "Why" :

I choose to contribute in this project because it meets my future goals, I have learned how data is processed, analyzed, and used in different systems in my IIT Madras program. I mean RAG first retrieves relevant information and then the LLM generates the answer using that information and who can know better about data and information than a Data scientist. I see too many connecting dots here and just wants to understand all at once(maybe I’m too curious) but this is not how it happens, so I’ll be breaking down everything and work on understanding them.

 3. My Goals :

Science I have the understood the concepts of SLM/LLMs and data processing through my Data Science program now my goal will be to learn about how RAG system works, especially Ramalama RAG and learn how I can use it to develop a SLM/LLM. I’ve watched some videos on the Red Hat Developers channel on youtube and got the roughly idea on how it works. Now I’ll try to connect as many dots possible to make a better understanding in the next few weeks.

 4. Just for fun:

If I could become an expert in something non-technical, it would be traveling. I enjoy exploring new places and cultures, I love just packing up few things and take off to a unknown place and be there to listen what it speaks and sometimes even without packing up something​.