Job description

Phiston Technologies builds high-security data destruction systems deployed at major data centers and Fortune 500 enterprises worldwide — ISO 9001/14001/45001 certified, NIST 800-88 and DIN 66399 compliant.

We're building automated vision-based asset identification into our equipment — systems that read, decode, and verify identifiers on media as it moves through destruction, producing the auditable records our clients rely on for compliance.

What you'll do:

Build and deploy computer vision/ML pipelines for barcode and label recognition (rotation-tolerant detection, multi-code disambiguation, OCR fallback)
Deploy models to edge compute on production hardware (quantization, ONNX/TensorRT)
Integrate vision output with machine controls (PLC/HMI, serial, Modbus, Ethernet/IP)
Build REST APIs connecting machine data to back-office/ERP systems for compliance reporting
Serve as technical owner for outside engineering vendors — scope work, review deliverables, keep quality in-house

What you bring:

4+ years building production embedded, machine vision, or applied ML software
Strong Python, working C/C++
Hands-on experience training/fine-tuning CV models (YOLO family or equivalent) — not just using pre-trained ones
Barcode/2D code decoding experience (ZBar, ZXing, libdmtx) — including where they fail
Experience deploying to edge/embedded compute (Jetson or similar)
Onsite in Miramar, FL, working directly with equipment on the floor
Bonus: OCR on industrial imagery, oriented object detection, PLC/controls integration, vision-guided robotics, vendor management experience.