Case study
Transforming search with Machine Learning
Founded in Copenhagen in 2019, Podimo is a platform for podcast and audiobook lovers as well as audio content creators. Their app puts a world of podcasts and audiobooks at your fingertips. Their revenue-sharing model rewards creators for making content that people want to hear – whether they have an established following or a niche audience.
Client
Podimo
Founded in Copenhagen in 2019, Podimo is a platform for podcast and audiobook lovers as well as audio content creators. Their app puts a world of podcasts and audiobooks at your fingertips. Their revenue-sharing model rewards creators for making content that people want to hear – whether they have an established following or a niche audience.
Partnership
Ongoing since November 2022
1000+
Exclusive Shows
70.000+
Audiobooks
Industry
Country
Project details
STX Next specialists were working together with Podimo onsite team to create an application that will allow users to listen and upload the podcasts and audiobooks. The app shares members’ subscription fees between the creators of the audio material so there would be a direct connection between a customer and a creator. Currently we are working on the bugfixing, feature updates and maintenance as the application is already live and working for the good of the customers.

The challenge
Podimo's Elasticsearch system delivered search results that weren't always relevant. Despite possessing rich information about user preferences, they struggled to consistently prioritize the most popular shows in search responses.
The solution
STX Next first scrutinized and optimized Podimo's existing Elasticsearch settings, refining mappings and analyzers. We then implemented a Learning To Rank machine learning model to reposition search results, placing those with higher historical user preferences at the top. Additionally, semantic search capabilities using BERT were integrated, enabling users to search using natural language and get more pertinent results. We're now in the process of adding a conversational search feature, primarily leveraging OpenAI's ChatGPT, allowing users to interact with a chatbot for their queries.






.webp)






.webp)


.webp)

.webp)


.webp)

.webp)

.webp)




































