Senior Applied Computer Vision Engineer

This is a fully remote position, open to applicants in Europe.

📋 Description

• Design and enhance computer vision models for sports video, focusing on player and ball detection, tracking, event recognition, and identity association.

• Develop and refine camera calibration, homography, and field-registration solutions that convert image coordinates into normalized field coordinates.

• Evaluate current computer vision pipelines, establish benchmarks, identify weaknesses, and suggest actionable improvements.

• Enhance tracking reliability across various stadiums, camera placements, broadcast styles, video qualities, and environmental conditions.

• Create experiments that encompass data acquisition, dataset creation, augmentation, model training, fine-tuning, evaluation, and readiness for deployment.

• Assess failure modes and implement enhancements that boost accuracy, reliability, scalability, and robustness.

• Modify existing models and pipelines to accommodate new sports, leagues, camera configurations, and video sources.

• Collaborate with data teams on labeling workflows, ensuring dataset quality, validation processes, and human-in-the-loop improvement cycles.

• Work closely with software, platform, and DevOps engineers to deploy computer vision models and pipelines in production settings.

• Optimize inference performance, scalability, monitoring, and operational reliability.

• Define evaluation metrics, testing protocols, and quality controls to ensure consistent model performance over time.

• Lead projects from initial technical exploration and prototyping to production deployment and ongoing enhancements.

• Contribute to system design decisions that integrate computer vision, machine learning, backend services, operations, and client workflows.

• Clearly communicate technical trade-offs with internal teams, client stakeholders, and engineering leadership.

⛳️ Requirements

• Extensive hands-on experience in building and enhancing production-grade computer vision systems.

• Proficient in Python and contemporary machine learning frameworks such as PyTorch.

• Experience with video-based computer vision challenges, including object detection, multi-object tracking, event recognition, identity association, or video analytics.

• Strong understanding of geometric computer vision concepts, including camera calibration, homography estimation, projective geometry, and linking image-space detections to real-world 2D or 3D coordinates.

• Experience in designing or refining tracking systems that address occlusions, object interactions, identity preservation, noisy detections, and missing information.

• Proven experience in evaluating model performance, identifying failure modes, and implementing practical enhancements.

• Background in adapting models to complex real-world data where video quality, camera angles, camera placements, and environmental conditions vary significantly.

• Knowledge of transfer learning, domain adaptation, data augmentation, and fine-tuning models on domain-specific datasets.

• Solid software engineering principles and the capability to write clean, maintainable, and production-quality code.

• Ability to work autonomously, prioritize tasks effectively, and drive technical initiatives to completion.

• Excellent communication skills and the ability to collaborate directly with clients and cross-functional engineering teams.

🏝️ Benefits

• Competitive salary with benefits, including paid vacation and sick leave.

• Opportunity to collaborate with a globally diverse team of top engineering talent on the industry's most challenging engineering problems.

• Highly flexible working conditions – we offer a generous office equipment allowance for remote work, the option of a desk at an office or coworking facility nearby, or a combination of both. No business travel required.

• A friendly start-up work environment, with exceptional opportunities for professional growth and development.

• Flexible working hours – as a remote-first company, we prioritize getting the job done well, rather than when or where it is completed.