- Company Name
- BOS Innovations Inc.
- Job Title
- AI Vision Programmer
- Job Description
-
**Job title**
AI Vision Programmer
**Role Summary**
Design, develop, and deploy deep‑learning computer vision models for industrial inspection and automation. Drive end‑to‑end model lifecycle from research to production, ensuring real‑time performance, high accuracy, and seamless integration into robotic and inspection systems.
**Expectations**
- Deliver production‑ready vision solutions that meet accuracy, speed, and reliability standards in manufacturing settings.
- Demonstrate continuous improvement of models through experimentation, synthetic data usage, and iterative tuning.
- Collaborate effectively with hardware, automation, and cross‑functional teams to integrate vision algorithms into robotic cells and inspection platforms.
**Key Responsibilities**
- Design, train, and fine‑tune models for defect detection, segmentation, classification, and 3D imaging using PyTorch or TensorFlow.
- Validate models on real and synthetic datasets, optimizing for generalization, latency, and resource efficiency.
- Deploy models into production pipelines (robotic cells, inspection systems, automation workstations) and support runtime monitoring.
- Build and automate synthetic data and labeling pipelines to augment training sets and reduce manual effort.
- Conduct lab‑based R&D, proof‑of‑concepts, and controlled experiments (e.g., synthetic vs. real data efficacy).
- Optimize AI/ML infrastructure, data storage, and training workflows for scalability and cost efficiency.
- Document model architectures, experiment results, best practices, and prepare technical reports for stakeholders.
**Required Skills**
- Proficiency in Python, PyTorch, and/or TensorFlow for deep learning model development.
- Strong understanding of computer vision tasks: object detection, instance segmentation, image classification, and depth estimation.
- Experience with synthetic data generation, data augmentation, and labeling pipeline automation.
- Familiarity with industrial vision hardware, robotic interfaces, and real‑time inference constraints.
- Ability to evaluate model performance metrics, conduct hyperparameter tuning, and implement model compression or acceleration techniques.
- Solid problem‑solving, analytical, and documentation skills.
**Required Education & Certifications**
- Bachelor’s degree (or higher) in Engineering, Computer Science, AI, Data Science, or related field.
- No mandatory certifications required; relevant courses in deep learning or computer vision are preferred.