Kate Rissman   •   8 months ago

Cheat sheet

Do you have any key advice or helpful insights that you would recommend from start to finish?

• 2 comments
• 

  Sara S Manager   •   8 months ago

  I'd recommend checking out the video that was just shared as well as the Resources page of the hackathon:
  https://drive.google.com/file/d/1vuYduwqyuI3Ept7985nTKw2f1D8JLX3P/view
  https://nvidia-aws.devpost.com/resources

• 

  Borna Rahnamay Farnod   •   8 months ago

  Hi Katie, there was just an office hour in the hackathon's Discord channel. I recommend that you join and read the conversation there. People from both NVIDIA and AWS shared some useful resources that can get you up to speed.

  Here's the link: https://discord.gg/kQujk5uy5J

  Once you join the devpost, you have to go hackathon checkins section and react to Sara's post (https://discord.com/channels/861966823054639134/1417543547581894816)

  I have participated in several hackathons and I personally want to prioritize 1. having a working code 2. presenting it for this hackathon. Since you will be sharing your code, it's very important that it works just outside your demo. And for presentation (visuals and the video) you wanna show how your agent operates, but also what is the impact of your work (e.g., how many hours saved?) and who can benefit from this. Even more, if you can show some market research that shows there's a need for this tool or even better have someone try it out and show it to the judges, that'll be very useful.

  From talking to one of the NVIDIA team members for this hackathon, one thing that they should see is that your agent has some sort of agency (logic/reasoning/planning/tool calling/ etc...) so that it can be considered as an Agentic AI. Here's more on Agentic AI: https://blogs.nvidia.com/blog/what-is-agentic-ai/

  These two points touches on two out of four criteria that the judges will be looking at. The other one is the user experience and also quality software development for which I recommend looking at some of the examples the team shared in Discord.

  - resource 1 (a video an NVIDIA member made for this hackathon): https://drive.google.com/file/d/1vuYduwqyuI3Ept7985nTKw2f1D8JLX3P/view?usp=sharing

  - resource 2 (example of deploying llama nemotron on sagemaker): We have an example notebook here of how to deploy the Llama nemotron model on sagemaker: https://github.com/NVIDIA/nim-deploy/blob/main/cloud-service-providers/aws/sagemaker/deployment_notebooks/nim_llama3.1-nemotron-nano-4b-v1.1.ipynb

  - resources 3 and 4 (blogs posts for deploying on EKS)
  - https://aws.amazon.com/blogs/hpc/deploying-generative-ai-applications-with-nvidia-nims-on-amazon-eks/
  - https://aws.amazon.com/blogs/hpc/deploying-generative-ai-applications-with-nvidia-nim-microservices-on-amazon-elastic-kubernetes-service-amazon-eks-part-2/

Comments are closed.