Complete TransPack Solutions is building a prototype mobile warehouse utility app that runs on-device AI to count stacked inventory in real time. This contract role is hybrid based in Fort Atkinson, WI (53538), with regular onsite collaboration to support hands-on prototype development and testing.
You’ll have the chance to train and deploy a custom object detection model for pallet and layer components, then integrate it into a mobile experience that overlays live bounding boxes and counting results. The project is estimated to take 4 to 8 weeks, with an opportunity to take on additional work once the prototype is complete.
Compensation includes an hourly wage of $30 to $55/hr based on experience (listed as $30.00 to $50.00 per hour) and expected 40 hours/week. Start date is available immediately. The role also includes a flexible schedule and a $5,000 completion bonus based on project completion.
Responsibilities
- Develop and prototype a mobile warehouse utility app that instantly counts stacked warehouse supplies using on-device AI.
- Train a lightweight custom object detection model (options include YOLOv8, YOLOv11, YOLO-NAS, or YOLOX) to recognize pallets, top frames, and layer pads across varying lighting conditions and angles.
- Set up a lightweight annotation workflow (for example Roboflow or CVAT) and prepare labeled data from 300–500 images of inventory.
- Export and optimize the trained model for local, offline inference on mobile using CoreML or TensorFlow Lite.
- Build a clean UI that integrates the model to show real-time counting overlays and bounding boxes on a live camera feed or uploaded photo.
- Co-work directly with one of the company owners to iterate on and test the prototype.
Requirements
- Authorization to work within the United States.
- Ability to work independently and research deployment bugs.
- Portfolio or GitHub link showing past projects where an object detection model was trained and successfully deployed to production or an edge device (examples listed: webcam, phone, Raspberry Pi).
Technologies
YOLOv8, YOLOv11, YOLO-NAS, YOLOX, Roboflow, CVAT, CoreML, TensorFlow Lite, Python, OpenCV, PyTorch, TensorFlow, Dart, Flutter, JavaScript, React Native, Swift, Kotlin, Apple Neural Engine, GPU, Raspberry Pi
Work Location and Onsite Collaboration
Hybrid remote in Fort Atkinson, WI 53538. The role is also described as seeking applicants based in the Adairsville area who can support frequent onsite collaboration.
- Regular travel between the plant in Adairsville for team co-working and hands-on prototype testing.
- Ideal candidates live within a reasonable commuting area.
