A leading U.S. drug testing network operator came to Armakuni for privacy-first IoT sensor system with AI/ML anomaly detection to prevent cheating in court-mandated urine drug tests across 120 testing centers.
The challenge
Scale: 120 testing centers, 200+ tests per center daily (24,000+ tests per day nationally).
What we built
Armakuni delivered privacy-first IoT sensor system with AI/ML anomaly detection to prevent cheating in court-mandated urine drug tests across 120 testing centers.
The outcomes
6 measurable outcomes shipped across the engagement. The ones that moved the business the most:
Armakuni created the dataset, the pipeline, and the working detection model that had never existed in this domain, and delivered a clear technical. path to reducing a $240,000 per-center annual monitoring cost.
First labeled sensor dataset for court-mandated drug test fraud detection.. No training data existed before Phase 1. Armakuni built it through controlled testing across the full range of known cheating methods. Every future model in this domain starts here.
HIPAA-compliant AWS pipeline at full production scale from day one.. KMS encryption, least-privilege IAM, versioning, and Lambda-driven indexing to DynamoDB. Built to handle 24,000 daily tests without architectural changes as volume grows.
Validated anomaly detection model with documented performance metrics.. Confusion matrices and ROC curves on held-out data. Operating thresholds tunable per site. Error analysis by cheating method and gender. False-alarm rates within agreed KPI targets.
End-to-end scoring demo running in a live environment.. Submit a session, receive an anomaly score and reason codes, log the result. Proven before any production deployment.
Pilot roadmap with success criteria, timeline, and ROM budget.. Recommendations across edge firmware hardening, AWS architecture, data management, and pipeline improvements for a subset-of-centers pilot deployment.
Datagain needed a team that had expertise across different domains and was enthusiastic enough to iterate on a moonshot project. [Armakuni] had a long term vision… I wanted people who were willing to say ‘yes’ and deal with the questions. Armakuni came highly recommended by AWS.
Vivek Jha, Managing Director, DIMS Product Suite, Datagain Services
Built on AWS
Delivered under our AWS Healthcare Consulting Competency, Data & Analytics Consulting Competency competencies.
