reelancer Job Posting: AWS ANPR Demo Environment Setup Title: AWS Real-Time Video Analyst Needed: Set Up DJI Drone RTMP to ANPR Demo Pipeline (Cost-Controlled) Category: Cloud Infrastructure / AWS / Machine Learning & Computer Vision Project Goal: We require an AWS expert to set up a functional, turn-key, and cost-efficient Automatic Number Plate Recognition (ANPR) demonstration environment on our AWS account. The solution must process a live video feed from a DJI drone streaming via RTMP and display the results in real-time. We have $3,000 in AWS credits to utilize for this demo environment setup and subsequent usage. Key Requirements & Architecture: The architecture must follow AWS best practices for real-time video analytics and adhere to a pay-as-you-go / minimal-idle-cost model. Ingestion Layer (Drone → AWS): Set up Amazon IVS (Interactive Video Service) or a cost-optimized EC2/NGINX-RTMP gateway to receive the live RTMP stream from the DJI drone. (IVS is preferred for simplicity if cost-effective). The stream must be reliably passed to Amazon Kinesis Video Streams (KVS). Processing Layer (ANPR Inference): Deploy a pre-trained, open-source ANPR model (e.g., a simple OpenALPR or public YOLO model) onto an Amazon SageMaker Endpoint. Implement the connection logic (e.g., using AWS Lambda or KVS Processors) to take frames from KVS, invoke the SageMaker Endpoint for ANPR, and collect the results. Visualization & Data Storage: Store the detected plate number, timestamp, and metadata in an Amazon DynamoDB table. Provide a basic, browser-based demo dashboard (e.g., hosted on S3/CloudFront or Amplify) that reads from DynamoDB and displays the ANPR results in near real-time. Critical Deliverables (Focus on Usability & Cost Control): The primary objective is to deliver a system we can easily operate for our demos without incurring unnecessary costs. Fully Functional AWS Environment: All resources (IAM roles, Security Groups, IVS/KVS, SageMaker, DynamoDB) configured using Infrastructure as Code (IaC) where possible (e.g., CloudFormation, Terraform, or CDK). Simple Start/Stop Scripts: Provide a clear, one-command script or documented process for the following: START DEMO: Quickly spins up the costly SageMaker Endpoint. STOP DEMO: Immediately and reliably shuts down the SageMaker Endpoint and any non-essential compute to prevent idle billing against our credits. Operation Guide: A simple, concise, step-by-step guide for non-technical users on: Where to find the drone's RTMP stream address. How to execute the START/STOP scripts. How to access the real-time results dashboard.