🌸 MaternaMind – AI Postpartum Diagnostic Assistant

Inspiration

Postpartum depression affects millions of mothers, yet screening typically occurs only during scheduled clinical visits. Many symptoms emerge at home, where support may be limited. We were inspired to build a solution that enables continuous, AI-assisted monitoring, helping detect risk earlier and provide structured support before symptoms escalate.

What it does

MaternaMind is an AI-powered postpartum mental health assistant that:

• Conducts a 20-question core assessment
  
• Tracks daily mood and sleep patterns
  
• Converts voice journals into structured symptom data
  
• Extracts diagnoses and medications from uploaded medical records
  
• Calculates dynamic risk levels (Low, Moderate, High, Critical)
  
• Generates clinician-ready summaries and patient-friendly guidance
  
• Triggers safety alerts when necessary
  

How we built it

We developed a multi-modal AI pipeline:

• Questionnaire and daily logs for structured inputs
  
• NLP for symptom extraction
  
• OCR for medical record parsing
  
• A rule-based risk stratification system combining emotional, behavioral, and historical indicators
  

Outputs are formatted into structured, EMR-ready summaries to support clinical workflows.

Challenges we ran into

• Designing a clinically meaningful risk model within limited time
  
• Balancing comprehensive screening with a simple user experience
  
• Integrating multiple AI components into one cohesive workflow
  
• Ensuring safety escalation without overwhelming users
  

Accomplishments that we're proud of

• Building a true RAG based diagnostic assistant, not just a questionnaire
  
• Implementing dynamic risk stratification
  
• Generating structured clinician summaries automatically
  
• Designing a safety-first, empathetic user experience
  

What we learned

• Continuous monitoring provides richer insight than episodic screening
  
• AI in healthcare must prioritize clarity, safety, and usability
  
• Structured outputs are critical for real-world clinical adoption
  

What's next for MaternaMind

• Training predictive models using real-world anonymized datasets
  
• Integrating directly with FHIR-based EMR systems
  
• Adding clinician dashboards for remote monitoring
  
• Conducting pilot testing with maternal health providers
  

MaternaMind’s next step is moving from prototype to real-world clinical validation, ensuring scalable and responsible deployment.

Built With

• django
• faiss
• llm
• rag