GAGANPREET SINGH
AI Product Manager | LLM Products (NL-to-SQL, Fitness SaaS) | Medical Imaging ML · 1st on BraTS-PEDs, MICCAI 2026 Leaderboard
- Handle
- @gaganpreet
- Role
- President
- Company
- NSUT Incubation and Innovation Foundation (verified: @nsut.ac.in)
- Location
- New Delhi
- Languages
- English, Hindi, French
- Open to
- Associate Product Manager, AI Product Manager, Applied AI / ML roles
About
I build AI products from problem discovery to deployment. My work spans LLM applications (a fitness SaaS and a production NL-to-SQL agent), medical-imaging deep learning, and explainable ML for public-sector users.
Currently ranked 1st on the BraTS-PEDs 2026 MICCAI Validation Leaderboard for pediatric brain tumor segmentation (DSC 0.950 Whole Tumor, 0.922 NET/C) using a two-model nnU-Net + MedNeXt ensemble. Earlier, I benchmarked 6 ML models on 4,500+ flood events and shipped a FastAPI + Streamlit interface for non-technical Central Water Commission stakeholders. As President of NSUT's Incubation & Innovation Foundation, I led 20+ summits with 50+ global leaders and 1,000+ participants.
I work across product discovery, experimentation, prompt engineering, product analytics (GA4, SQL) and end-to-end delivery. Seeking Associate / AI Product Manager and applied AI roles.
Experience
- AI Intern at Center of Excellence in AI
Jun 2026 to Jul 2026 - New Delhi
1st on the BraTS-PEDs 2026 MICCAI Validation Leaderboard; first author of a paper accepted at MICCAI 2026 and invited to present the work in Strasbourg, France. Developed a two-track ensemble using nnU-Net and MedNeXt for 4-class pediatric brain tumor segmentation. - ML Research Intern at NextGenAI Lab
May 2025 to Jul 2025 - New Delhi
Built an ensemble deep-learning pipeline using 3-D CNNs for 4-class tumor response classification on longitudinal MRI data, improving macro AUROC on the imbalanced Lumiere clinical dataset via class-weighted sampling, hybrid loss, and model ensembling. - President at NSUT Incubation and Innovation Foundation
Sep 2024 to now - New Delhi
Led 10+ entrepreneurship summits and speaker series featuring 50+ global leaders, attracting 1, 000+ participants. - Summer Intern at Central Water Commission
Jun 2024 to Jul 2024 - New Delhi
Engineered 50+ features across 5 hydrological datasets covering 4, 500+ flood events for end-to-end flood-level forecasting. Benchmarked 6 ML models, with XGBoost achieving R2 = 98.5%, RMSE = 25.47, and MAE = 4.38.
Education
- Netaji Subhas University of Technology, Delhi
B.Tech, CSE
2023 to 2027
CGPA: 7.74/10 - S. S. Mota Singh Model Sr. Secondary School
Board (Class XII), Computer Science, Physics, Chemistry, Mathematics
2021 to 2022
77.8% - Guru Amar Das Public School
Board (Class X), Information Technology
2019 to 2020
84.5%
Skills
- LLMs
- Generative AI
- Prompt Engineering
- AI Prototyping
- Product Discovery
- Experimentation
- Product Analytics
- GA4
- SQL
- Python
- Deep Learning
- Machine Learning
- Medical Image Segmentation
- Stakeholder Management
- C++
- NLP
- Data Science
- ML
- DL
- PyTorch
Links
This profile as JSON