Case Study: Media Publishing Company informs voters with an NLP platform built by Wizeline

Building a NLP Platform to Keep US Voters Informed of Presidential Candidates’ Stances on Trending Topics

Media Publishing Company (MPC) partnered with Wizeline and AWS to build a public NLP platform that could answer users’ natural-language questions about U.S. presidential candidates’ positions on trending topics. The challenge was to search decades of debate and appearance transcripts, surface relevant quotes quickly, and let users filter results by candidate, issue, date, or keyword.

Wizeline implemented a web platform using AWS services such as Amazon API Gateway, Lambda, DynamoDB, S3, Kinesis Firehose, Amazon Kendra, and SageMaker, along with MPC’s proprietary ingestion tools. The solution ingested and processed transcripts, extracted key phrases, topic-tagged content, and indexed it for fast search; it launched in September 2020 and attracted 40k+ users in its first two weeks. Wizeline also introduced QA and DevOps best practices, helping MPC improve software quality and adopt a more secure development cycle.