Emerging Five: Building a Real-Time Brand Logo Detection System
Built a production-grade logo detection pipeline and stakeholder analytics dashboard during a one-month internship.
Alqama Ansari
Quantitative Researcher at Confidential ( Stealth Mode )
From their time as
AIML Intern
Emerging Five • 2025 - 2025
Overview
During his internship at Emerging Five, Alqama was tasked with building a real-time brand logo identification system. The goal was to give stakeholders a reliable way to measure advertisement exposure and campaign ROI across a large volume of marketing content.
The Story
During his internship at Emerging Five, Alqama was tasked with building a real-time brand logo identification system. The goal was to give stakeholders a reliable way to measure advertisement exposure and campaign ROI across a large volume of marketing content.
He built the detection pipeline using a YOLOv8 object detection architecture, achieving 95% accuracy on the identification task. Alongside the core model, he developed an interactive Streamlit dashboard with confidence-scoring algorithms, giving stakeholders a direct interface to run advertisement exposure analytics and measure campaign performance.
He also designed a scalable framework to calculate logo display percentages across more than 50 distinct marketing campaigns, establishing a benchmarking system for ongoing campaign performance measurement.
This was his first significant exposure to stakeholder-facing work. He ran regular meetings with the client to understand their requirements, reported on project progress, and incorporated their feedback into the build. The experience gave him an early foundation in translating a technical deliverable into something a non-technical audience could use and act on.