Independent research portfolio exploring AI-assisted pediatric vision screening, real-world validation, and healthcare AI deployment challenges.
This repository supports my independent research interest in AI-assisted pediatric vision screening, healthcare AI deployment, and real-world validation challenges in early childhood eye care. The goal of this repository is to document learning, literature review notes, public resources, technical experiments, and responsible AI considerations related to pediatric vision screening.
Pediatric vision conditions such as amblyopia and strabismus can have long-term consequences when detection and intervention are delayed. Recent advances in artificial intelligence, computer vision, and digital health have created new opportunities to explore more accessible screening support systems.
However, important challenges remain around:
- Real-world validation
- Dataset diversity and generalizability
- Accessibility in community settings
- Clinical workflow integration
- Responsible and equitable healthcare AI deployment
This repository is intended to serve as a public research portfolio, including:
- Literature review summaries
- Publicly available resources
- Healthcare AI notes
- Computer vision learning experiments
- Responsible AI and validation considerations
- Project updates related to pediatric vision AI research
This repository now focuses on responsible AI-assisted early pediatric vision risk detection before age 3, with emphasis on hospital and pediatric healthcare workflows, EHR-connected risk signals, referral pathways, and real-world validation.
The goal is not clinical diagnosis but to explore how AI can support earlier identification of children who may need professional eye care evaluation.
Current focus:
- Literature review
- Domain understanding
- Public dataset exploration
- Healthcare AI validation concepts
- Responsible AI considerations
- Medium article: https://medium.com/@nipashah2311/why-pediatric-vision-ai-still-faces-real-world-challenges-cde1906376c3
- LinkedIn article: https://www.linkedin.com/pulse/100-million-children-have-lazy-eye-crossed-eyes-ai-can-nipa-s-7rmpe/
- Community awareness article: https://patch.com/new-jersey/jersey-city/your-child-could-have-vision-problem-right-now-what-every-jersey-city-parent-nodx
Responsible AI-Assisted Early Pediatric Vision Risk Detection Before Age 3
Published: July 2026 DOI: https://doi.org/10.5281/zenodo.21386465 Access: Open Access
- Completed CITI Program Biomedical Research Refresher 1, covering human subjects research protections, informed consent, IRB awareness, privacy, and responsible conduct considerations relevant to healthcare AI research
- https://www.citiprogram.org/verify/?w57352384-4e66-4ee7-b4e0-b2b159dc681f-77378035
- CITI Program — Artificial Intelligence (AI) and Human Subject Protections
- https://www.citiprogram.org/verify/?wec03ac2c-68fc-4a90-bec8-479e2661de66-77378040
Nipa Shah
Senior Data Scientist | Independent Researcher
MS Business Analytics — Sacred Heart University
MBA
LinkedIn: https://www.linkedin.com/in/nipa-s-486287382/ ORCID: https://orcid.org/0009-0009-4115-9652
This repository represents independent research exploration by Nipa Shah. All written research concepts, workflow frameworks, responsible AI documentation, article summaries, and original analysis are authored by Nipa Shah unless otherwise cited.
This work is shared publicly for educational, research, and collaboration purposes. Reuse, adaptation, or citation should provide appropriate attribution.
Commercial use, clinical deployment, medical device development, or proprietary use of this work requires written permission from the author.
This repository reflects independent research exploration and educational work. It is not intended to provide medical advice, clinical diagnosis, or clinical guidance.
I welcome thoughtful discussion with researchers, clinicians, healthcare AI professionals, and public health professionals working at the intersection of pediatric vision care, medical imaging, and responsible AI.