MongoDB has introduced ORiGAMi, a Transformer-based architecture aimed at simplifying machine learning on semi-structured data like JSON without the need for conversion to tabular form, addressing a key challenge faced by data scientists. This innovative model, developed by MongoDB's ML research group, tokenizes documents into sequences of key-value pairs and structural tokens, allowing efficient training on datasets with as few as 200 labeled samples. ORiGAMi's approach enables direct predictions from semi-structured documents while preserving their hierarchical relationships, thus streamlining processes like user segmentation. Meanwhile, ZEE5, a leading OTT video platform, successfully migrated its operations to MongoDB Atlas, achieving zero downtime during the transition of 40+ microservices, showcasing MongoDB Atlas's capabilities in handling high traffic and complex workflows. Additionally, MongoDB is undergoing a leadership transition, with CJ Desai set to succeed Dev Ittycheria as CEO, bringing extensive experience from ServiceNow and Cloudflare to guide the company through its next evolution.