Computer Vision Engineer Technical Pre-Sales & Delivery

As a Computer Vision Engineer at TVARIT GmbH, your role will involve bridging the gap between business requirements and technical execution. You will work on various real-world Computer Vision (CV) use cases, from pre-sales to deployment, and will be responsible for technical feasibility and delivery. Your responsibilities will include: - **Technical Pre-Sales & Solutioning**: - Engage with customers to understand problem statements, process flow, constraints, and success criteria. - Translate business requirements into clear CV solution approaches. - Perform feasibility analysis covering camera placement, optics, lighting, FPS, accuracy expectations, and infra constraints. - Propose end-to-end CV architectures including cameras, edge/cloud compute, models, and dashboards. - Support pre-sales discussions, demos, and technical presentations with clients. - Define scope boundaries, assumptions, risks, and dependencies for CV projects. - Work closely with sales/product teams to ensure realistic commitments on accuracy, timelines, and effort. - **Core Computer Vision Engineering**: - Handle dataset preparation: data cleaning, annotations, QA, and format conversions (COCO, YOLO, VOC). - Train and fine-tune models using Ultralytics YOLO, Detectron2, MMDetection with appropriate augmentations and hyperparameter tuning. - Build inference pipelines for images and videos (frame extraction, RTSP streams, batch processing). - Apply classical image processing techniques using OpenCV and NumPy where deep learning is not sufficient. - Export and optimize models using ONNX / TensorRT for deployment on NVIDIA GPUs or edge devices. - Support deployment, testing, and iterative improvements based on field feedback. - **Visualization, Demos & Tooling**: - Build quick PoC dashboards or demo UIs using Dash/Plotly or simple GUIs (PyQt/Tkinter) for client demos. - Create visualizations for inspection results, trends, defect distributions, and analytics. - Support internal and customer-facing demos with clean, understandable outputs. **Must Have Skills**: - Strong Python skills with hands-on experience in PyTorch-based CV frameworks. - Proven experience delivering multiple real-world computer vision use cases. - Strong fundamentals in image processing, camera optics, lighting, and video analytics. - Experience taking CV projects from idea to deployment. - Comfortable interacting with customers, operations teams, and non-technical stakeholders. - Ability to clearly explain technical trade-offs, limitations, and risks. **Preferred Skills**: - Prior experience in technical pre-sales, solution consulting, or customer-facing engineering roles. - Ability to estimate effort, timelines, and risks for CV projects. - Experience writing or contributing to technical proposals, scope documents, or solution notes. - Exposure to industrial environments. - Comfortable proposing alternatives early. **Bonus Skills**: - Hands-on experience with NVIDIA DeepStream for real-time pipelines. - Experience building Dash/Plotly dashboards or lightweight analytics tools. - Familiarity with FastAPI / backend services for CV deployments. - Exposure to edge devices. - Experience with annotation tools like CVAT, Label Studio, Roboflow. Join TVARIT GmbH and be part of a dynamic team driving advancements in Predictive Quality, Predictive Maintenance, and Energy Consumption Reduction through cutting-edge AI solutions for the metal industry. As a Computer Vision Engineer at TVARIT GmbH, your role will involve bridging the gap between business requirements and technical execution. You will work on various real-world Computer Vision (CV) use cases, from pre-sales to deployment, and will be responsible for technical feasibility and delivery. Your responsibilities will include: - **Technical Pre-Sales & Solutioning**: - Engage with customers to understand problem statements, process flow, constraints, and success criteria. - Translate business requirements into clear CV solution approaches. - Perform feasibility analysis covering camera placement, optics, lighting, FPS, accuracy expectations, and infra constraints. - Propose end-to-end CV architectures including cameras, edge/cloud compute, models, and dashboards. - Support pre-sales discussions, demos, and technical presentations with clients. - Define scope boundaries, assumptions, risks, and dependencies for CV projects. - Work closely with sales/product teams to ensure realistic commitments on accuracy, timelines, and effort. - **Core Computer Vision Engineering**: - Handle dataset preparation: data cleaning, annotations, QA, and format conversions (COCO, YOLO, VOC). - Train and fine-tune models using Ultralytics YOLO, Detectron2, MMDetection with appropriate augmentations and hyperparameter tuning. - Build inference pipelines for images and videos (frame extraction, RTSP streams, batch processing). - Apply classical image processing techniques using OpenC