Personalied Recommendation Maker

THE CUSTOMER

Personalized Recommendation Maker is a Real-time Personalization Platform that utilizes Augmented Intelligence to help enterprises transform customer experiences. Personalized Recommendation Maker is a Machine Learning (ML) Personalization Platform born out of Google’s ML Sibyl Research team. It helps customers personalize their products and services to improve core business metrics such as user engagement, conversions, and revenue. It uses cutting-edge ML techniques.

THE PROBLEM

Personalized Recommendation Maker aims to optimize and streamline its ML tools to generate accurate and quick recommendations .
Personalized Recommendation Maker needed enhancements and support to maintain the system to support quick data processing.
It also was to be ensured that the bandwidth consumption using the GCP was optimized to prevent overload.
Realtime Activity was needed for the recommendations generation process based on catalog scraping to generate recommendations based on cookies and browsing history.
Accessibility is important considering the presence of users across the globe and in different regions, hence instant results are expected at different locations simultaneously.
Huge database storing more than 100 M items in the global database makes for a knotty situation to be able to read and process the data and generate recommendations.

FOLIO3 SOLUTION

For in-time live dashboarding, and to view the performance across multiple regions and locations where Personalized Recommendation Maker is being used, Global public load balancers have been implemented and maintained for balancing across regions.
S3 Bucket Storage has been used and maintained for efficient storage of objects on the system.
Managed Instance groups are used to operate multiple VMs on the Google Cloud making the workload scalable and highly available with its autoscaling, auto-healing, regional deployment, and automatic updating features.
Google Cloud SQL is used as a relational database, as it provides simple and efficient migration of data to Cloud SQL from MySQL, PostgreSQL, SQL Server, and Oracle databases.
Google Cloud API Gateway is used as an option to create, secure, and monitor APIs for Google Cloud serverless back ends. This allowed for reduced complexity, security and visibility.

Folio3 Helps Personalized Recommendation Maker in managing GCP infrastructure and manage codebase in the following ways:

Process website activity data by creating Apache Spark compute instances, loading activity data from Google Cloud Storage, loading Inventory Items from AeroSpike DB, generate performance reports using that data.
Investigate issues reported by Personalized Recommendation Maker's clients, up scaling and down scaling instance groups, manage clusters, mainly increasing/decreasing nodes depending on need and restarting.
Using GCP SQL and BigQuery to investigate activity data and generate performance reports for clients, BigQuery reports are near real time as it is also used in the dashboard to report all kinds of performance metrics to XYZ Company's clients.
Develop new features for their console dashboard based ReactJs.
Monitor infrastructure via GCP Logging for any important failures during the night shift.
Fixing Build pipelines running on VMs or compute instances running Jenkins, often fixing disk usage issues by increasing Cloud Storage Size.
Investigates recommendation quality problems and fixes them by tweaking model weights, occasionally implementing a new feature to resolve the report issue.
Investigates any model server infrastructure alerts and failures on Google Cloud Computes and VMs.
Heavily rely on Google Cloud BigQuery to investigate runtime alerts by making frequent queries to BigQuery and generating reports on demand.
Analyze activity and performance data on Prometheus Monitor.
Performs software versions upgrades, including, driver migration, ReactJs upgrade, NodeJs upgrade, and others.
Maintain and upgrade the codebase.
Helps Personalized Recommendation Maker customers debug script, rendering, tracking errors.

Personalized Recommendation Maker

Personalized Recommendations using ML

THE CUSTOMER

THE PROBLEM

Timesheets

Approvals

Procurement

FOLIO3 SOLUTION

Folio3 Helps Personalized Recommendation Maker in managing GCP infrastructure and manage codebase in the following ways:

Technologies Involved in Case Study

Get In Touch

Contact Us

Contact Info

Email Address

Phone Number

Follow Us

US OFFICE

UAE Office

UK OFFICE

Mexico Office

PAKISTAN OFFICE

CANADA OFFICE

Bulgaria Office

SummitK12

Moodle Based LMS

Optimizely

A Leading AB Testing Platform