Computer Vision Performance Analyst

Role Overview

We are looking for a detail driven analyst to support SeeChange teams to evaluate the performance in live deployments of its computer vision system for selfcheckout (SCO). The system monitors checkout activity to detect missed scans, hidden items, and items left behind after the transaction ends, prompting shoppers to correct the issue or, when needed, triggering an employee intervention driven analyst to checkout (SCO). The system monitors checkout activity to detect missed scans, hidden items, and items left behind after the transaction ends, prompting shoppers to correct the issue or, when needed, triggering an employee intervention.

Your primary responsibility will be to review annotated SCO video clips, verify the ground truth labels, and assess whether the detections are accurate. You will ensure that alerts retailers consider valuable significantly outweigh those they would regard as ‘noise’, helping protect both operational efficiency and the customer experience.

The insights you produce will directly support SeeChange teams in delivering implementations across new clients, retail formats, and use cases. They will shape how performance is communicated to clients, inform decisions on model retraining, alert threshold adjustments, and rule development, and directly influence both the success of customer projects and the prioritisation of our development roadmap. A key part of this role is serving as the super user and administrator of the video review system managing user onboarding and maintaining a strong understanding of the ingestion pipelines and the end to end data flow process.

Key Responsibilities

Solution performance QA
Review annotated SCO camera footage to validate events such as:
• Missed scan attempts
• Item concealment/hidden items
• Items left in the output tray after transaction
• Incorrect bagging behaviours

Confirm timestamps, bounding boxes, and event labels against ground truth:
• Flag incorrect labels, inconsistencies, or unclear cases
• Maintain a clear categorisation of SCO events to ensure reliable evaluation.

Model Performance Evaluation
Measure and interpret key CV performance metrics per SCO event type:
• Precision, recall
• False positive and false negative rates

Operational & Customer Experience Alignment
Ensure false alerts do not outweigh genuine detections to preserve:
• Smooth customer flow
• Shopper trust
• Staff workload balance

Data & Quality Governance
Working together with the data analyst ensure that video samples are representative of real SCO traffic (quiet vs. peak, basket sizes, lighting conditions).

Video Review System Governance
Define and implement the categorisation of video clips, ensure consistent classification, and maintain high QA standards while shaping how insights and conclusions are presented so they can be easily consumed and used to support the effective deployment of SeeChange solutions.

Skills & Experience Required
• Experience in video review, annotation QA, or computer vision evaluation.
• Understanding of common SCO behaviours and loss prevention risk points.
• Ability to analyse CV metrics and translate them into practical insights.
• High attention to detail for repetitive but highly impactful review work.
• Competence with Excel/Sheets or BI tools for tracking performance over time.
• Strong communication skills with the ability to simplify technical findings.

Nice to Have
• Experience with retail SCO or loss prevention.
• Understanding of SCO hardware setups (scanners, cameras, bagging areas).
• Knowledge of operational metrics: intervention rate, throughput, basket size distribution.

Key Performance Indicators (KPIs)
• Positive stakeholder feedback from Engineering, Data Science, Delivery, and New Business teams
• Clear, well‑structured documentation and high‑quality insight reports
• Ability to quickly understand expected solution performance and identify when system behaviour falls outside defined parameters
• Proactive oversight and governance of the video‑performance review system
• Effective communication with team members and reliable delivery against agreed deadlines