engblogs

Explore how Databricks Lakehouse architecture dispels common data modeling myths by integrating robust relational principles, constraint enforcement, and advanced transaction capabilities to deliver a unified, high-performance platform for modern data warehousing and analytics.

BBVA and OpenAI partner to deploy ChatGPT Enterprise across 120,000 employees, driving an AI-native transformation in global banking to enhance customer experience, optimize operations, and innovate financial services.

BNY scales enterprise AI with a platform-first, governance-driven approach using Eliza and OpenAI to responsibly deploy 125+ live use cases and digital agents across 20,000 employees.

Leveraging the GPT-5.1-Codex model, a lean engineering team rapidly developed and launched the high-quality, crash-free Sora Android app in 28 days by combining AI-assisted coding with rigorous architectural oversight and contextual guidance.

BNY implemented a centralized AI platform, Eliza, integrating OpenAI models with robust governance and extensive employee training to democratize AI use across the enterprise and transform banking operations responsibly and securely.

A novel method from MIT researchers improves the accuracy and reliability of confidence intervals in spatial statistical estimations by accounting for spatial smoothness and data bias.

MIT's new 2N6 program combines mechanical engineering and AI education to train naval officers with advanced applied AI skills for enhanced military decision-making and national security.

MIT researchers developed DisCIPL, a framework where large language models coordinate smaller models to efficiently solve complex reasoning tasks with high accuracy and reduced cost.

AI-assisted configuration-driven SQL dialect generation combined with automated 25,000-query validation accelerated the expansion of Zero Copy connectors in Data 360, enabling scalable, reliable pushdown to 100+ external SQL engines with 4x faster development throughput.

Explore the advancements in Gemini audio models enhancing voice technology for more powerful and immersive user experiences.

NeurIPS 2025 retrospective forecasts 2026 as the year of environments, highlighting data-centric AI, principled evaluation, and scalable benchmarks for LLM agents across diverse tasks and environments.

GPT-5.2 delivers state-of-the-art long-context reasoning, robust agentic tool-calling, and professional-grade capabilities across coding, spreadsheets, and presentations, driving new benchmarks like GDPval and SWE-Bench to new heights.