AI Platform for Precision Medicine

Research project developing a comprehensive platform for analyzing clinical data of brain tumor patients through Artificial Intelligence algorithms, enabling personalized diagnosis and classification.

My Role

As a researcher and developer, I focused on:

  • 🏗️ Development and deployment of the complete platform
  • 🧠 Research and development of specialized AI models
  • 🎯 Training Deep Learning models for medical analysis
  • 🔬 Scientific validation of algorithms and results

Key Features

  • Multimodal analysis: Integration of clinical data, images, and omics data
  • Medical Deep Learning: Specialized models for histological and radiological images
  • Omics processing: Analysis of transcriptomics, epigenomics, and oncological genomics
  • Assisted diagnosis: Automated brain tumor classification
  • Clinical interface: Intuitive system for use by medical professionals

Tech Stack

Backend & ML

  • PyTorch + Lightning: Main framework for Deep Learning models
  • FastAPI: High-performance REST API
  • PostgreSQL: Database for clinical data storage

Frontend

  • React: Modern and responsive web interface

Infrastructure

  • Distributed training pipeline
  • Model validation and evaluation system
  • Sensitive medical dataset management

Technical Challenges

  • Multimodality: Effective integration of heterogeneous data (images, text, omics data)
  • Limited data: Data augmentation and transfer learning strategies for small datasets
  • Medical privacy: Compliance with healthcare data protection regulations
  • Explainability: Implementation of interpretability techniques for clinical decisions
  • Clinical validation: Metrics adapted to real medical context

Impact

Platform that enables doctors and researchers to:

  • Accelerate brain tumor diagnosis
  • Identify new therapeutic biomarkers
  • Personalize treatments according to patient profile
  • Facilitate precision medicine research

Funding

Project funded by the Valencian Innovation Agency (AVI) under file INNEST/2022/159, in collaboration with ISABIAL, FISABIO, AIJU, and The New Horizons.