Innovation at the Edge: Inside the Qualcomm compute AI hackathon in NYC, in collaboration with LM Studio and Microsoft
Introduction
Qualcomm, in collaboration with LM Studio, Microsoft, and NYU Tandon school of engineering, hosted an AI on the Edge hackathon at the NYU engineering campus in Brooklyn on Dec 7th and 8th. The event challenged AI developers to build real-world edge Large Language Model (LLM) applications with tangible commercial value using only the power of the Snapdragon X Elite NPU and LM Studio.
Hackathon participants
Out of more than 160 prospective registrants, 45 applications were selected yielding 20 driven, focused teams – and did they deliver! Working tirelessly through the night, every team managed to not only complete the application, but they also prepared and delivered product pitches highlighting their product, how they harnessed the NPU, and the challenges they overcame along the way. Some of the standout projects include:
|
R.E.D.A.C.T. |
Tool for reading and removing PII from documents |
|
|
Legal Doc Analyzer |
Automated contract analysis tool |
|
|
Policy Pal |
Helps people figure out their insurance issues - and minimizes the work to get insurance providers to pay them! |
|
|
Tutor AI |
Local LLM to assist teachers in rural areas, addressing challenges like language diversity, limited resources, and varied literacy levels among students. The LLM should support multiple local dialects, provide instant translations. |
|
|
JobLinkers |
A local AI powered tool that helps job search, customize resumes and cover letters and automatically fills in all the data for you. |
|
|
Mindful |
Personal mental health companion |
https://github.com/mannadamay12/MindfulCompanion/tree/develop |
|
MeWayWise |
Travel mode recommendation tool for inclusivity |
|
|
DevDuck |
Debugging tool companion for coders |
|
|
SnapLearn |
Learning companion for students |
|
|
VS Code Security |
VS extension for code vulnerability scanning |
These projects were well-built and well-pitched, and we encourage you to check out their code bases and demos.
If you would like to see in-depth demos and tutorials of these tools, reach out to us on Discord to tell us which teams we should ask to present.
Overall, the event was a resounding success for everyone involved. Participants shared positive feedback about the tools, software, and hardware provided. The NPU was the standout performer, garnering rave reviews in the feedback portion of the demos:
“The NPUs are really good for inferencing, and they are super power efficient. For example, even when working during this hackathon I just lost 15% of battery while I had the LLM hosted for 5 hours and an online meeting for one hour. That is so great!” - Satyam Chatrola (check out his Github!)
Participants appreciated the support from Qualcomm, LM Studio, and Microsoft experts throughout the event, and were thrilled to keep the Snapdragon laptops upon completion of their projects.
To participate in future events, connect with us on Discord and be sure to sign up for the Qualcomm Developer newsletter to be notified of our upcoming developer events across the Unites States, Asia, and Europe.
