Founding Senior Computer Vision Engineer

About the role

A client of Flatgigs, which builds AI match analysis for youth football, is hiring a founding senior computer vision engineer. Fixed cameras at the pitch record every session, and the models detect passes, shots and goals, track players and read jersey numbers from low-resolution footage. The product is already running at a pilot academy in Dubai and is being expanded to a second industry. The hire owns the shared CV platform across both, improving the live sports models and building the new product on the same stack.

Key responsibilities

• Own the shared CV platform end to end: camera ingestion, detection and tracking, OCR, classification and the reporting pipeline, powering a product in production and a second one launching.
• Ship to a live product from week one: raise accuracy on sports models with real users and architect the second industry's deployment on the same stack.
• Define architecture: model selection and training strategy, annotation and dataset operations, evaluation methodology, and the on-prem/cloud split.
• Deploy the first releases for the new industry and harden both products as one platform.
• Lead the engineering team you will build.

Required profile

• 5+ years building computer vision systems that ran in production, with deployed systems and users, not notebooks.
• Experience building datasets, running annotation operations and measuring models against ground truth.
• Has fine-tuned open-source models (detection, OCR such as PaddleOCR/TrOCR) and replaced them with custom models trained on validated annotations, with an annotation-review loop.
• Fluent in the modern CV tooling ecosystem: Roboflow or equivalent for datasets and annotation, experiment tracking and model registries.
• Ownership temperament: turns a vague operational goal into a working system without waiting for a spec.

Required skills

• Strong Python and PyTorch, hands-on with modern detection and tracking (YOLO family, ByteTrack/DeepSORT or similar).
• Video pipeline: RTSP/IP cameras, scheduled or streaming inference, ffmpeg-level fluency.
• Pluses: OCR, action recognition or pose estimation in production; sports, retail, industrial or facility camera analytics background; edge deployment (Jetson) and cost engineering; camera and IoT literacy (sensor/lens selection, PoE, ONVIF/RTSP, NVRs).

Hiring process

• Intro conversation about a CV system you built, with decisions, failures and numbers.
• Technical deep-dive with senior engineers.
• Case study on real footage: spec and prototype a detection task.
• Founders' conversation, then offer; the whole process within two weeks.

The role is full-time and remote-friendly, with working hours overlapping the Gulf and travel to Dubai for installation and milestone weeks. The package is described as competitive plus options.