Computer Vision Expert for Coral Ecosystem Monitoring

Project Overview We are seeking an experienced Computer Vision specialist to architect and build future-proof solutions for our coral reef restoration and aquaculture operations. The ideal candidate will help us develop a unified approach to processing underwater imagery and extract valuable data that improves our operations. About Us CoralVita is focused on coral reef restoration, in ocean coral nurseries, and land based coral aquaculture. We use imaging technology to monitor coral growth, health, and restoration success. We need expert help to scale and improve our computer vision capabilities. Key Responsibilities - Architect a comprehensive computer vision framework that addresses our immediate needs while laying groundwork for future applications - Develop unified detection and segmentation models to avoid a library of many hyper specialized models - Improve, expand on Batch ID detection systems, currently using QR codes - Create solutions for automated photomosaic processing for: - Benthic classification (identifying and classifying the type of ocean floor e.g. coral, rock, sand, seagrass, etc) - Coral count and size estimation - Coral health/bleaching classification - Genus and Species identification Current Technical Challenges 1. Image Processing Pipeline Improvement - Batch ID recognition accuracy - Processing underwater photomosaics (large orthomosaics, stitched together composits of thousands of photos) - Standardizing our segmentation approach across different imaging contexts 2. Coral Health Monitoring - Color analysis to detect bleaching specimens - Currently we use k-means clustering but are open to other solutions - Detecting additional pre-mortality markers 3. Growth Tracking - Identifying individual coral specimens across timepoints - Measuring vertical and horizontal growth of coral specimens Required Skills & Experience - Strong background in computer vision and deep learning - Experience with object detection, image segmentation, and classification models (bonus for knowledge and expertise in cutting edge transformer architectures and photogrametry approaches like Structure from Motion and NERF) - Experience in un-supervised computer vision like k-means clustering - Familiarity with modern computer vision infastructure and frameworks. - Currently we use and have access to Databricks and Roboflow. While we are open to alternatives, these are where our current implementations live. - We have significant credits on GCP and access to high power VMs as well as Vertex AI - Ability to critically evaluate which problems need immediate solutions and advise on long-term architecture decisions - Proven experience building production-ready computer vision pipelines Nice to Have - Experience with underwater imagery or environmental monitoring - Knowledge of GCP's Vertex AI or similar cloud ML platforms - Background in biological image analysis - Experience with 3D model segmentation Project Timeline & Engagement This is initially a short-term consulting engagement with potential for ongoing collaboration. We're looking for someone who can: 1. Evaluate our current approach and technical debt 2. Recommend a cohesive strategy that prioritizes high-impact solutions 3. Build a proof-of-concept for our most pressing issues 4. Outline a roadmap for future development If you're passionate about using computer vision to support environmental conservation and have experience building practical, production-ready solutions, we'd love to hear from you!