Why Every CS Student Should Contribute to Open Source

Yash Saraogi
Yash Saraogi
Cover Image

For decades, edge computing and AI were seen as separate areas of innovation. But in 2025, something exciting is happening — generative AI models are becoming small enough to run directly on edge devices.

And this change presents a huge opportunity for ACM student chapters to take the lead.

Whether you’re into embedded systems, machine learning, or open-source, the intersection of LLMs and edge computing is the perfect playground for ACM developers and enthusiasts.

Why Should ACM Care About Edge AI?

ACM is all about advancing computing as a science and profession. With AI at the edge, we’re now seeing:

  • Local LLMs that summarize documents offline
  • AI voice assistants working without cloud connections
  • Real-time translation and computer vision in robotics and AR

These breakthroughs directly align with ACM’s focus on innovation, ethics, and building for impact.

What’s Driving This Shift in 2025

LLMs are getting smaller and smarter
Models like Phi-3 Mini, OpenELM, and Gemma 2B are optimized for edge deployment.

Hardware is catching up
Devices like NVIDIA Jetson Nano, Raspberry Pi 5, and even Android flagship chips can now run transformer models.

Open-source tools are widely available
Libraries like GGML, Llama.cpp, and ONNX Runtime make deployment fast and student-friendly.

Ideal ACM Club Projects to Try

Here are a few edge-AI project ideas your ACM chapter can build, demo, or publish as research:

  • AI-powered Campus Guide
    A wearable or phone app that uses a local LLM to answer campus-related queries offline.
  • Code Autocomplete on Raspberry Pi
    Use a distilled code LLM to help students write Python code on a low-cost device.
  • Offline Chatbot for Mental Health
    Privacy-focused LLM chatbot trained on wellness resources — works without internet.
  • Generative Art or Story Creator for Events
    Let attendees generate short stories or AI art on-site using local models.

These projects are not only technically impressive but also great for ACM paper submissions or inter-college hackathons.

Learning Resources for ACM Students

  • fast.ai, HuggingFace Courses — For ML model fine-tuning
  • Edge Impulse, tinyML Foundation — For embedded AI
  • Google Colab + Llama.cpp — To test quantized models before edge deployment
  • ACM Digital Library — For access to related research papers

Beyond Code — Ethics and Inclusion

As part of ACM’s mission, any work in generative AI should also consider:

  • Data bias in training datasets
  • Security of offline AI systems
  • Accessibility and multilingual support

Your ACM club can even host panels, debates, or workshops around these issues.

Conclusion

Generative AI at the edge is more than a tech trend — it’s a movement. As AI becomes more compact and accessible, ACM chapters across campuses have the chance to lead real change, develop new solutions, and push the limits of student innovation.

If your club is looking for a 2025 tech theme that combines software, hardware, ethics, and impact — this is it.