Clinical Training & Patient Simulation

DokSim

Advanced patient communication simulator powered by Python 3.11, LangGraph agentic state machines, and Chroma vector stores. Allows medical students and residents to conduct virtual patient interviews with real-time empathy rubrics and diagnostic evaluation.

Python 3.11LangGraphChroma DB RAGFastAPI

Institutional Pilot Inquiry

Real-time Patient EvaluationFastAPI + LangGraph

Safe Branching Dialogue

LangGraph agentic state machines maintain safe patient boundaries while dynamically responding to emotional cues, body language probes, and clinical line-of-questioning.

Real-Time Evaluation Rubrics

Calculates instant feedback on clinical rapport, diagnostic specificity, empathy scores, and protocol adherence following standard OSCE medical guidelines.

Medical Vector Retrieval

Integrates Chroma DB vector search across peer-reviewed clinical knowledge bases, ensuring patient personas output grounded medical facts without hallucinations.

Interactive Simulation

DokSim Virtual Patient Interview

Select a clinical patient case and choose interview probes to observe real-time AI evaluation.

Eleanor Vance (Age 64)OSCE CASE

Chief Complaint: Acute Chest Tightness & Dyspnea

Persona traits: Anxious, hesitant to share medication history

Marcus Brody (Age 42)OSCE CASE

Chief Complaint: Severe Epigastric Pain & Nausea

Persona traits: Stoic, minimizes symptoms, high stress job

Select Clinical Question Probe:

Or type custom question:

PATIENT_AGENT // Eleanor Vance (Age 64)BRANCH: Symptom Onset Explored

Patient Verbal Response:

“It started about 2 hours ago while I was gardening. At first I thought it was just heartburn, but then I felt winded...”

Empathy & Rapport Score:

94 / 100

Diagnostic Signal:

Symptom Onset Explored

Clinical Evaluator Feedback:

Good open-ended inquiry. Patient revealed onset context and respiratory involvement.

Python 3.11 / LangGraph EngineResponse Latency: 32ms