Company
Swathmark is developing a computer-vision detect-and-decide platform for visible, threshold-managed specialty-crop pests.
Mission
Build the reusable decision engine that turns operational field imagery into local treatment decisions that can be defended — evidence in, accountable action out.
Detect. Decide. Protect.
Why the decision layer matters
Finding a pest in an image is one step of nine. Swathmark is built around the full chain — because the chain, not the detection, is what makes a field action defensible.
Operational capture
Rights & provenance
Data-quality verification
Pest & life-stage detection
Density & biological context
Versioned threshold policy
Treat / no-treat / abstain
Machine-readable output
Decision traceability
First proving domain
A visible, biologically staged, threshold-managed pest — the hardest honest test of a capture-to-decision loop, and a concrete proving ground for a reusable engine.
The data, policy, evaluation, and equipment interfaces are designed to extend across additional visible, threshold-managed pests, specialty crops, regions, capture systems, and equipment platforms.
Operating principles
A treatment decision carries its full chain of evidence: source media, rights status, model version, calibration, threshold policy, and review history.
Dedicated treatment machinery is outside the current platform scope. Equipment integration begins from a defined software contract, not from building a machine first.
No performance, savings, or adoption claims without an approved evidence record. Product previews are labeled for exactly what they are.
Training eligibility is enforceable in the platform itself — recorded, versioned, and auditable — never an undocumented note.
Leadership
Founder & Chief Executive Officer
Aadya leads Swathmark's strategy, product vision, and partnership initiatives. She holds a BS from the University of Michigan and an MS from the University of Maryland, and founded the company on a platform-first conviction: agriculture will be shaped by reusable perception engines that equipment makers can integrate — not by isolated, crop-specific point products.
Chief Technology Officer
Peter leads technology strategy, platform architecture, and the engineering organization. With an MS from the University of Maryland and more than two decades building enterprise software and production AI systems, he architects the vision engine for edge-native inference on working equipment — with modular APIs and SDKs for OEM integration, and an insistence that every decision the system produces is explainable and reproducible.
Chief Research Engineer
Andrew leads the research behind Swathmark's vision engine, focused on the hard problem of perception that generalizes across real field variability — light, weather, biology, and terrain. He holds an MS in Computer Science specializing in AI, computer vision, and high-performance computing, and drives the GPU-accelerated training infrastructure and model-efficiency work that lets sophisticated networks run on constrained field hardware.
Working on specialty-crop protection, field capture, entomology, or application equipment? Let's talk.
