Welcome to the Operative Intelligence Platform Engineering Assessment. We thank you in advance for taking the time to complete this.
This assessment will introduce a hypothetical scenario to you, your task is to read through the scenario below and start thinking about how you would engineer a solution to fit. You don't need to present anything to us, we will schedule a technical interview with you to discuss your approach and thinking.
The technical interview will be a video call with a couple of our engineers here at OI. We will ask you a few questions related to the scenario below to help better understand your thinking and approach in the following topics:
- Networking
- Security
- Monitoring and Observability
- Site Reliability
- Scalability
During the interview we encourage you to dive into technical details, explain your thought process and consider trade-offs.
Intelligent Operations Pty Ltd is launching a new product called Eye-Q. It is a real time crowd intelligence platform used to calculate wait times in crowded events such as large sporting venues, concerts, music festivals and theme parks (similar to https://www.thewaittimes.com/).
The platform consists of:
- A video capture service which consumes video streams for thousands of IP camera's
- A crowd detection service which uses a pre-trained ML model to detect crowd sizes in realtime from a video stream
- A data ingestion and preprocessing service which collects data from the crowd detection service then processes the data for analysis
- A cluster of Postgres databases which stores historical and real-time data for live and future analysis
- An API service responsible for reporting and analysis. This service generates reports on crowd movements and incidents, it also allows for in-depth analysis of historical data
- A reporting dashboard that provides an interface for data visualisation and viewing of live video streams.
- A model training service used by data scientists and ML engineers to train models for the crowd detection service
Some important things to consider:
- Each cluster of IP camera's exist in separated co-located networks
- While some processing may occur on a customer premise it is expected that all the data eventually will end up in Intelligent Operations environment
- Privacy is a big concern as Intelligent Operations aims to remain compliant and secure.
- Data sovereignty laws may come into play when processing data for government clients. This means that their data cannot leave their respective region.
- Intelligent Operations is operating at very strict SLA's, with an expectancy of 99.99% uptime across all global regions
- When developing for the platform, it is important for the developers at Intelligent Operations to move at a fast, yet stable pace