About the role
Our client specializes in AI match analysis for youth football. Utilizing fixed cameras placed at the pitch, they record every session, detecting key events such as passes, shots, and goals, while tracking players and reading jersey numbers even from low-resolution footage. This innovative approach provides academies and parents with match statistics, highlights, and per-player records. Currently operational at a pilot academy in Dubai with actual matches and paying pilots, the company is expanding its technology into a second industry, with initial customers already committed. This role will take ownership of the platform across both industries by enhancing the live sports models and developing a new product using the same foundational technology.
Ready to apply for roles like this?
Unlock the company name and direct application link. Subscribers get instant access to fresh jobs across Dubai, Abu Dhabi and Riyadh, many with visa support.
Unlock employer & apply directly
What you own
- The shared CV platform, covering end-to-end processes: camera ingestion, detection and tracking, OCR, classification, and reporting pipeline, which powers a product already in production and a second one on the way.
- Immediate contributions to a live product from week one, improving the accuracy of sports models with real users while also architecting the deployment for the second industry on the same technological stack.
- Responsibilities regarding architecture, including model selection and training strategy, annotation and dataset operations, evaluation methodology, and the on-prem/cloud split.
- Shipping the initial deployments for the new industry and subsequently hardening both products into a unified platform.
- Providing technical leadership to the engineering team that will be built under your guidance.
Hiring process
- Introductory conversation: Discuss your experience in building a CV system, including decisions made, failures encountered, and performance metrics.
- Technical deep-dive with senior engineers.
- Case study involving real footage: Specifying and prototyping a detection task.
- Final conversation with the founders, followed by an offer.
- The entire process is expected to be completed within two weeks for the right candidate.
What we need to see
- At least 5 years of experience in building computer vision systems that have been deployed in production environments, with direct user interaction rather than theoretical notebooks.
- Strong proficiency in Python and PyTorch, with hands-on experience in modern detection and tracking technologies (e.g., YOLO family, ByteTrack/DeepSORT).
- Experience with video pipelines, including handling RTSP/IP cameras and knowledge of scheduled or streaming inference, as well as fluency with ffmpeg-level tasks.
- Demonstrated experience in building datasets, managing annotation operations, and measuring model performance against ground truth honestly.
- Experience in fine-tuning open-source models (detection and OCR, specifically PaddleOCR/TrOCR class) and advancing them to the next level by replacing them with custom models trained on validated annotations, including active involvement in a working annotation-review loop.
- Familiarity with contemporary CV tools like Roboflow or similar for dataset and annotation operations, experiment tracking, and model registries.
- An ownership mentality, with the ability to transform a vague operational goal into a functioning system without needing detailed specifications.
Strong pluses
- Experience with OCR, action recognition, or pose estimation in production settings.
- Background in camera analytics in sports, retail, industrial, or facilities.
- Knowledge of edge deployment solutions (e.g., Jetson) and cost-effective inference engineering.
- Literacy regarding cameras and IoT, including sensor and lens selection, PoE networking, ONVIF/RTSP understanding, NVRs, and weatherproof housings; capable of site specification beyond just model performance.
Contract and benefits
- Full-time position with remote work flexibility, aiming to hire the best qualified candidates regardless of location.
- Working hours should overlap with the Gulf time zone, including travel to Dubai for installations and key milestone weeks.
- Competitive compensation package along with options.
