Real-Time Diagnostics from Nanopore DNA Sequencers
Sepsis is a dangerous condition when the body attacks its organs and tissues to fight bacteria. This process is fast and needs a quick solution -- which antibiotic to use?
The University of Queensland researchers and the Google Cloud Platform developers cooperated with the APP Solutions to create a tool for doctors to save patients from septic shock.
Such a tool needs a vast amount of data processing; thus, significant resources. The APP Solutions’ team managed
to develop a cloud-based solution to this challenge. This time, we were engaged with machine learning, data analysis, and Google Cloud platform capabilities.
This system is dedicated to navigating and operating within a vast amount of various information in real-time. The product aims to make money by providing a robust platform for ads designed for mobile app installations.
Our client was expanding their business, so they needed to scale an existing CPA product. The system was supposed to handle dramatically higher levels of traffic. So, our client requested us to develop a custom affiliate marketing system, aimed at mobile conversions. We had to expand and improve the data processing toolset — make it scalable, add new parameters and metrics, create an intuitive design.
Our team revised, upgraded, and moved the project to the cloud. The product was about to be complex, so we opted for Google Cloud services. The backend was optimized to process 2,000 queries/sec on average (with the peak load around 5,000 queries/sec).
After all, we provided the customer with a platform for registering new
CPA targeting systems. The analytics include over 50 parameters with 20 metrics to describe every impression, click, and conversion.
Fraud in online ads might become a big problem. The APP Solutions created a
custom ad fraud detection system designed to identify and report malicious adverts before they cause damage.
The main challenges for the team were scalability, improvement of the fraud detection algorithm, and development of the crawler engine. Major quantities of information to be processed and the necessity to scale the project cried out for a cloud solution. Thatis why we cooperated with Google Cloud.
Our team studied fraudulent content and researched how it usually gets around the protection.
In this way, we extended the scope of our fraud detection mechanisms which allowed us to develop an efficient fraud detection system that can work properly under a significant workload.