Monarch Waste Case Study
From Clipboards to Clean Data: Digitizing Pharmaceutical Waste Destruction for Monarch Waste Technologies
How Pravaah replaced paper logs and gut-feel judgment with a connected data platform that traces every waste box from the dock to the machine, and laid the groundwork for AI-driven blending.
Company Overview
Monarch Waste Technologies (MWT) is a U.S.-based waste technology company that specializes in the secure destruction of medical, pharmaceutical, and other regulated products. Founded in 2015, MWT developed PyroMed, a proprietary non-incineration pyrolysis technology built to permanently destroy approved waste while cutting emissions and reducing the transportation that traditional incineration demands.
PyroMed earned an EPA determination that it is not an incinerator, a milestone that followed years of federal testing. Today MWT serves pharmaceutical manufacturers, healthcare organizations, biotech companies, medical-device manufacturers, and reverse distributors, all of whom care about the same things: compliance, chain-of-custody tracking, documented destruction, and sustainability.
Introduction
Inside a pyrolysis system, balance is everything.
The machine runs on a delicate thermal equilibrium. Wet waste, meaning anything with excess liquid, quenches the system's temperature. Dry waste such as pills, plastics, and cardboard keeps the heat steady. The skill of the operation lies in blending the two in the right proportions before each Gaylord (the large container that feeds the system) is tipped in.
Until now, that judgment lived in people's heads and on paper. Operators looked at each box, decided whether it was wet or dry, and worked out blending ratios by hand. The workflow was manual and paper-based, and the records it produced were unstructured, scattered, and impossible to learn from.
That created a ceiling. MWT wanted to move toward machine learning models that could predict the best blend and, eventually, automate it. But you cannot train a model on handwritten notes. The company needed a clean, structured, image-rich data foundation first.
That is where Pravaah came in.
Solution
Pravaah designed Phase 1 as the digital backbone for everything MWT wants to build next. The goal was a single, unbroken chain of data from the moment a box arrives to the moment the machine finishes with it.
The platform brings six capabilities together:
Digitized box intake and staging. A web application replaces handwritten logs. Operators capture high-resolution images of each box's contents along with the key details: Box ID, customer, operator name, and notes.
OCR and packaging classification. An automated OCR pipeline reads packaging slips and extracts the waste type, wet-or-dry classification, weights, and contents lists. When a slip is missing, a manual inspection path keeps the process moving.
Image labeling and dataset creation. Captured images flow into an annotation pipeline built on Label Studio, where subject matter experts label containers, blister packs, and packaging types. This builds the ground-truth dataset that Phase 2 AI training depends on.
Gaylord mix tracking. Each mix is recorded by linking the contributing Box IDs and weights into a new Gaylord record, backed by visual documentation taken during loading.
Machine telemetry integration. Through PLC and API connections, the platform automatically captures machine performance data, including BTU output, temperature curves, pressure readings, and cycle duration, and maps it to the right Gaylord ID.
End-to-end traceability. Everything connects in one structured chain: Box, then Mix, then Gaylord, then Machine Output. This is the foundation that makes AI predictions possible, and eventually automated blending and closed-loop optimization.
Action
Pravaah began with Phase 1 development, focused on capturing incoming waste boxes and waste Gaylords directly in the system so that data collection started at the very first touchpoint.
From there, the team:
Captured and annotated images throughout the waste mixing stage, so the visual record grows with every Gaylord prepared.
Configured Label Studio to receive, process, and route every captured image for systematic labeling by subject matter experts.
Digitized scheduled checklists, so tracking and record-keeping now sit alongside image capture instead of in a separate paper trail.
The approach was deliberate. Rather than jumping straight to AI, we built the data infrastructure that makes good AI possible, one module at a time, with each stage feeding the next.
Project Timeline
Phase 1 was delivered in four milestone waves over roughly ten weeks, from environment setup on May 6 to production launch on July 15.
Delivery Date : May 6
Modules : M1 Environment Setup, M2 Discovery and Design, M3 Box Intake Module
Delivery Date : Jun 6
Modules : M4 Manual Content Entry, M5 Image Review, M6 Label Studio and Annotation
Delivery Date : Jul 6
Modules : M7 OCR and Packaging Slip, M8 Gaylord Mix and Tracking, M9 BTU and PLC Integration
Delivery Date : Jul 15
Modules : M10 Analytics Dashboard, M11 User Acceptance Testing, M12 Production Launch
The sequence moved from foundations, to data capture, to intelligence, to launch, so the client could see working software at every stage.
Results
Phase 1 delivery is complete.
MWT now tracks incoming waste digitally, monitors the thermal processes inside its machines, and logs key operational parameters in real time. Handwritten logs and unstructured records have been replaced by a connected, searchable chain of data that links every box to its mix, its Gaylord, and the machine output that followed.
Just as important is what this unlocks. With images captured, labeled, and tied to machine telemetry, MWT now holds the structured dataset that Phase 2 needs: AI models that can predict the right blend, and over time, automated blending and closed-loop optimization.
For a company whose reputation rests on defensible, documented destruction, the paper trail is now a data trail.
