Role: AI Vision Lead— Edge & Multimodal AI
Business Unit: AI & Cybersecurity Practice
Location: (Onsite) WFO
Role Overview
We are seeking a hands-on AI Vision Lead to drive the design, development, and deployment of computer vision capabilities for industrial safety and operational intelligence solutions.
This role combines technical leadership with strong execution ownership, covering the full lifecycle of AI vision systems — from data strategy and model development to real-time deployment on edge platforms and continuous improvement in live production environments.
Key Responsibilities
- Lead development of computer vision solutions for industrial monitoring, safety intelligence, and inspection automation
- Define dataset strategy, annotation workflows, validation approaches, and performance benchmarks
- Architect and optimize deployment of real-time AI models on edge hardware (e.g., NVIDIA Jetson or similar platforms)
- Ensure reliability, scalability, and latency optimization for multi-camera video analytics systems
- Integrate perception outputs with enterprise applications, alert engines, and operational workflows
- Monitor and improve model performance in production through systematic error analysis, dataset refinement, and retraining cycles
- Evaluate and gradually adopt modern multimodal AI approaches such as video event understanding, vision-language reasoning, synthetic data usage, and sensor fusion
- Mentor engineers and collaborate with customers and stakeholders to translate operational challenges into scalable AI solutions
Required Skills & Experience
- 3 to 5 years of experience in computer vision or applied AI system development in production environments
- Strong hands-on expertise in Python and deep learning frameworks (PyTorch / TensorFlow)
- Experience building object detection or video analytics solutions
- Exposure to deploying optimized models for real-time inference (ONNX / TensorRT / Edge AI preferred)
- Experience managing training data workflows including annotation coordination, label quality validation, and dataset versioning
- Familiarity with Linux environments, Docker, and video processing pipelines
- Strong understanding of model evaluation metrics, reliability considerations, and real-world validation
Preferred
- Experience with edge AI platforms and multi-stream video inference optimization
- Awareness or exposure to multimodal AI concepts such as vision-language models or temporal video intelligence
- Experience in industrial, manufacturing, or safety-critical AI deployments
- Experience handling challenging visual conditions such as fire/smoke detection and false-positive reduction in complex environments
- Familiarity with model lifecycle management or MLOps practices
What Success Looks Like
- Robust AI vision capabilities deployed reliably across industrial environments
- Continuous improvement in detection accuracy, system performance, and scalability
- Clear evolution path toward next-generation multimodal AI vision systems
