Project: NeoSmartML Foundation Model for Neonatal ICU Monitor Data

Hi @r0bby , I hope you’re doing well.

I wanted to ask if there is any specific proposal template or format that LibreHealth expects us to follow for the GSoC proposal or should we follow the general GSoC guidelines?

I just want to make sure my proposal aligns with the expected structure.

Thanks in advance!

Yeah, it’s listed both on our gsoc profile and among the project list.

Also: Open Source, Open Doors, Apply Now for Google Summer of Code! | Google Open Source Blog

1 Like

Hi @r0bby , @sunbiz and @shbucher ,

I have completed my drafted proposal for the GSoC 2026 for “NeoSmartML Foundation Model for Neonatal ICU Monitor Data” project. I would really appreciate your feedback before I submit it in GSoC.

I have attached my proposal below:

NeoSmartMl GSoC Proposal.pdf (353.4 KB)

Could you please review this and suggest technical approach and changes which needed to be done before final submission.

Best regards,

R Sagar Shresti

Go ahead and submit it. It looks good, but I’m not making the final decision.

1 Like

Thankyou so much @r0bby for reviewing my proposal! I really appreciate your feedback and guidance.

Best regards,

R Sagar Shresti

I just glanced at it, I didn’t review it in detail…I just made sure it had a plan…That’s it :slight_smile:

1 Like

Thanks @r0bby , I will wait for @sunbiz and @shbucher to review it in detail.

I do have few questions regarding the project :

should contributors propose publicly available datasets for training or is their any specific dataset available to proceed with?

Is this project intended for production deployment in actual neonatal ICU environments or is it primarily for research?

should I start with ResNet50 or a similar framework to establish baseline performance then move to more optimized frameworks for edge deployment?

These clarifications will help me refine my proposal further.

Best regards,

R Sagar Shresti

Hi @r0bby, @sunbiz, @shbucher,

I have drafted my proposal for GSoC 2026 on the project NeoSmartML Foundation Model for Neonatal ICU Monitor Data and would really appreciate your feedback before final submission.

I’ve attached the proposal below:

NeoSmartML GSoC 2026 Anagha Pradeep.pdf (180.2 KB)

Please review it and share suggestions.

Your feedback would be extremely valuable in refining the proposal.

Best regards, Anagha Pradeep

1 Like

You used an LLM almost exclusively (I stated not to) and even not withstanding that your proposal lacks any detail. At this stage, it’s impossible to turn-around a proposal that removes all LLM-generated text. If submitted, it will be not considered.

Better luck next year and hopefully you don’t use an LLM then. I stated over and over not to do it.

Hey @r0bby, I want to make it clear that I didn’t use an LLM to create the proposal. I know it doesn’t have enough information, and I’m willing to start again and write it again before the 31st. Are you willing to look over a new version?

Don’t worry about grammar and spelling. No LLM use. No grammar, no spell check, nothing.

Here’s the report:

GPTZero AI Scan - NeoSmartML GSoC 2026 Anagha Pradeep.pdf.pdf (52.5 KB)

You need more detail, a lot more detail.

Thank you for the feedback and for taking the time to review it. I hear your concerns, and I respect them. I am willing to rewrite the entire proposal before the 31st without any grammar tools or spell check. The technical work is mine, and I can defend every decision in it. I just want the chance to show that.

Hi @r0bby, I have written the document without even using any grammar correction tool like quillbot this time. Please review it at your convenience.

GSoC 2026 Proposal.pdf (232.7 KB)

Thank you.

  1. Still uses AI – or at least leans on it. Ditch the template. Anything that you didn’t write, ditch it. Don’t paraphrase, ditch it.
  2. You should probably know the organization name you’re applying to – you’re showing us you don’t have good attention to detail

I’m not even looking at the rest until I see a better proposal.

Pay attention to our AI rules — absolutely NO LLM USE.

Your project plan and timeline are what we care about.