Real-Time Object Detection with Pravega and Flink

on YouTube ↗15m

" summary

A live pipeline ingests vehicle video frames from Jupyter into Pravega via a Pravega gRPC connector, runs a Flink job that applies a YOLO model using TensorFlow Java bindings to draw bounding boxes and append detection metadata, then writes enriched frames back to Pravega and tiers them to Dell EMC ObjectScale on vSphere with Tanzu. The full stack runs on Dell Technologies Cloud Platform with Flink clusters deployed from the SDP UI, and the demo shows live detection working end to end with continuous playback.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 0:38Dell Technologies Cloud Platform
3. 2:14Architectural View
4. 2:58Review the Infrastructure Deployment in Vsphere
5. 4:29Streaming Data Platform Architecture
6. 8:27Real-Time Object Detection Using Streaming Data Platform
7. 10:14Demo
8. 10:55Create a Flink Cluster
9. 11:25Flink Application

" related talks

• 14mLive Streaming Object Detection by Leveraging GPU with AVA Platform on EWAOLShuang Hu · SOAFEE Virtual Symposium 2022
• 53mVideo Object Detection at ScaleReza Zadeh · O'Reilly Artificial Intelligence Conference 2017 - New York, New York
• 25mReal-Time Fraud Detection at ScaleSatyanandam Kotha · Conf42 Machine Learning 2025
• 41mAdding Generative AI to Real Time Streaming PipelinesTimothy J Spann · Build Stuff 2024
• 53mVideo Analytics for mere mortals with NVIDIA DeepStream SDKYousry Mohamed · NDC Melbourne 2022
• 41mPredictive Maintenance with Apache FlinkDongwon Kim · Flink Forward Berlin 2017
• 18mCluster Serving: Distributed Model Inference using Apache Flink in Analytics ZooJiaming Song, Dongjie Shi · Flink Forward Global Virtual 2020
• 21mReal-time Stream Analytics and Scoring Using Apache Flink, Druid & CassandraCiesielczyk, Zontek · Flink Forward Europe 2019