Skip to content

Kantaka Śodhana

Kantaka Sodhana is an AI and MLOps platform for healthcare insurance fraud detection. Built for the National Health Authority of India, our system detects forged medical documents, radiology report mismatches, duplicate claims, same-day discharge fraud, and family clustering anomalies in Ayushman Bharat PM-JAY insurance claims. Winner at the NHA AB PM-JAY Auto-Adjudication Hackathon 2026 at IISc Bengaluru. Technologies include PyTorch, FastAPI, Next.js, computer vision, NLP, OCR, SHA-256 document verification, and deep learning models for deepfake detection.

AI & MLOps Platform

Kantaka Sodhana — AI-Powered Healthcare Fraud Detection Platform

Every fraudster leaves a fingerprint — in the data, the timing, the pattern.
We built the scanner.

scroll

Faster model delivery with automation

60%

Reduction in production failures

24/7

Monitoring + drift detection

100%

Traceability: data → model → deployment

·Airflow·Kafka·DVC·MLflow·PyTorch·Ray Tune·KServe·BentoML·Kubernetes·Prometheus·Grafana·Airflow·Kafka·DVC·MLflow·PyTorch·Ray Tune·KServe·BentoML·Kubernetes·Prometheus·Grafana
Kantaka Śodhana

Every fraud leaves a trace.
We find it.

In classical statecraft, Kantaka Śodhana meant the proactive removal of thorns — antisocial actors preying on the marketplace. We run that operation at machine speed.