Mayank verma

Senior Software Engineer

Amagi Media Labs•  September 2021 - Present

Project - Knowledge graph for asset metadata (Global Metastore) Python | SparQL | XML | XSL | RDF | Argo Contributed in building python based CLI tools and workflows for ingesting asset's metadata into a global metastore knowledge graph which extracts data from publisher's xml MRSS, transforms into a internal standard RSS using XSLs and then converts this standard RSS into RDF format to ingest into global metastore which uses AWS Neptune . Also wrote SparQL queries to fetch data from this DB to be consumed by other applications. Project- Third Party Partner API Integrations Python | FastAPI | MongoDB | Redis Developed APIs (based on FastApi) to facilitate our third party partners to insert ad markers (dynamic brand insertion and contextual video) in our video assets. Additionally, built a robust Test Automation Framework using Pytest, covering 80-85% of test cases for this API service and setup prometheus metrics for monitoring the service for any anomalies. Project - Speech to text Conversion Workflows Python | FastAPI | MongoDB | Argo Workflows Developed Argo cron workflows which generate subtitles for video assets which do not have srt files associated with them. Used google's speech-to-text model underneath for subtitle generation from extracted audio. Leveraged ffmpeg for audio extraction from video assets