Each colour maps a published value to the exact source passage used to support it. Only bounded excerpts are public; administrators can inspect the complete captured source.
Title
MediaRadar | Vivvix uses GenAI and computer vision on Databricks to automate video ad classification
…immediate challenges but also positioned them for future growth and innovation. The implementation of GenAI exemplifies their forward-thinking approach, enabling MediaRadar | Vivvix to serve ad insights faster and with more accurac…
Description
MediaRadar | Vivvix built a GenAI pipeline on Databricks using Spark Structured Streaming, the Ray cluster, Whisper for transcription/translation and OCR to classify video ads across millions of product categories. A dual-layer approach combines GenAI product identification with in-house classification models, increasing hourly ad-classification throughput from about 800 to 2,000 creatives and cutting model-experimentation time from two days to half a day.
…r Databricks. “Before, we were able to do at most around 800 creatives an hour. With our move to Databricks, we’re classifying about 2,000 an hour.” This remarkable increase in throughput underscored the platform’s ability to h…
Company
MediaRadar | Vivvix
…ights at scale. Manual data processing impedes valuable ad insights for clients MediaRadar | Vivvix, an advertising intelligence company, helps brand marketers and advertising agencies understand their competitive landscape and the impact of their media spends. Their primary mission is to discern “what…
Industry
Technology & Software
…e more effective decisions, visit vivvix.com/contact-us Share this post Details Industry : Technology and Software Use Case : Artificial Intelligence , Data Engineering Cloud : Azure Product : A…
Problem
MediaRadar | Vivvix faced manual data processing and fragmented workflows that couldn't keep pace with rapidly growing ad volumes. Their prior Amazon SQS setup, limited to manual polling of 10 messages at a time, hindered SLA compliance, and classifying over 6 million unique products was an extreme classification problem that existing in-house ML models and hundreds of human operators manually watching ads could not scale to match nearly doubling ad spend.
…y’s ability to meet service level agreements (SLAs). The nature of their task — classifying over 6 million unique products — posed “an extreme classification problem,” as Dong-Hwi Kim, MediaRadar | Vivvix’s Senior Machine Learning (ML) Engineer, p…
Solution
MediaRadar | Vivvix adopted Databricks and Apache Spark Structured Streaming for automated, continuous real-time data ingestion, using a Ray cluster on Databricks to scale video-ad classification. A dual-layer approach uses GenAI to identify products in ads and compares results with in-house classification models to select the best match. Preprocessing pipelines include fingerprinting for duplicate detection, Whisper for transcription/translation, and OCR for text extraction, with Databricks Model Serving used for rapid prototyping and OpenAI's GPT-3.5 chosen to balance performance and cost; the team is also moving to Databricks Unity Catalog for governance.
…ad classification process, significantly enhancing efficiency and scalability. The MediaRadar | Vivvix team leveraged the Ray cluster on Databricks to optimize video processing and scale the classification of video ads across millions of categories. This innovative, dual-layer approach involved using GenAI to identify products…
Business value
MediaRadar | Vivvix increased ad classification throughput from about 800 to roughly 2,000 creatives an hour, a 150% increase in ads categorized per hour, while model experimentation and testing time that used to take about two days now takes as little as four hours, and moving into Databricks eliminated the data silos and fragmented pod monitoring that previously made operations hard to track.
…r Databricks. “Before, we were able to do at most around 800 creatives an hour. With our move to Databricks, we’re classifying about 2,000 an hour.” This remarkable increase in throughput underscored the platform’s ability to h…
AI capabilities
Generative AI, Computer Vision, Document Intelligence
…ds across millions of categories. This innovative, dual-layer approach involved using GenAI to identify products in ads and then comparing the results with MediaRadar | Vivvix’s own classification models to select the best matches, ensuring higher accuracy in identifying the correct…
Technology
Databricks, Databricks Model Serving, Unity Catalog, Whisper, GPT-3.5
…ed preprocessing pipelines that included fingerprinting to identify duplicates, transcription and translation using models like Whisper and optical character recognition (OCR) to extract textual information from the ads. These preprocessing steps were crucial for preparing the data for…
Deployment model
Cloud
…: Technology and Software Use Case : Artificial Intelligence , Data Engineering Cloud : Azure Product : Agent Bricks Ready to get started? Try Databricks for free Learn more…
Deployment options
cloud
…: Technology and Software Use Case : Artificial Intelligence , Data Engineering Cloud : Azure Product : Agent Bricks Ready to get started? Try Databricks for free Learn more…
Headline outcome
…ents 2 days to 4 hours Time saved experimenting with Agent Bricks AI Playground 150% Increase in ads categorized per hour MediaRadar | Vivvix provides comprehensive advertising intelligence that empowe…
Use case type
Document processing
…s and Apache Spark™ Structured Streaming, they revolutionized their operations, automating ad classification and streamlining data management through a unified platform. This transformation enabled MediaRadar | Vivvix to process thousands of ads pe…
