IndiaMedia and entertainment2022 – ongoingDream11
Machine learning platform for personalization and fraud detection
Dream11 cut the time to launch new ML features from 5 months to 4 weeks on a central Databricks platform
Problem
Dream11 runs a fantasy sports platform with more than 120 million users. It needed to ship personalized recommendations and real-time predictions fast, and to handle traffic peaks during events such as the Indian Premier League.
Approach
Dream11 used Databricks on AWS as a central machine learning platform. It supported personalized and social recommendations, real-time predictions, fraud and anomaly detection and operational efficiency. A key personalization use case for the Indian Premier League was delivered within one week.
Results
Takeaway
A shared ML platform shortened the path from idea to launch. The company gave speed figures but no figures on user engagement or revenue.
Categories
- Industry
- Media and entertainment
- Business function
- IT and R&DMarketing and sales
- Innovation type
- Product and service innovation
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks reports that Dream11 uses its platform on AWS for personalization, recommendations, real-time predictions and fraud detection. Time-to-market for new ML solutions fell from 5 months to 4 weeks. Concurrency grew from 3 million to 5.5 million users and traffic from 40 million to 100 million requests per minute between 2020 and 2021.
Full details are in the original sources.
