In this week’s real-time analytics news: Researchers have developed a new approach to train open-source models.

Keeping pace with news and developments in the real-time analytics and AI market can be a daunting task. Fortunately, we have you covered with a summary of the items our staff comes across each week. And if you prefer it in your inbox, sign up here!

Researchers at Penn Engineering at the University of Pennsylvania and the Allen Institute for AI (Ai2) have developed a new approach to train open-source models: using AI to create scientific figures, charts, and tables that teach other AI systems how to interpret complex visual information.

Their tool, CoSyn (short for Code-Guided Synthesis), taps open-source AI models’ coding skills to render text-rich images and generate relevant questions and answers, giving other AI systems the data they need to learn how to “see” and understand scientific figures.

The resulting dataset, called CoSyn-400K, includes more than 400,000 synthetic images and 2.7 million sets of corresponding instructions, in categories as varied as scientific charts, chemical structures, and user-interface screenshots. CoSyn-trained models outperformed top proprietary systems like GPT-4V and Gemini 1.5 Flash on a suite of seven benchmark tests.

Real-time analytics news in brief

The Apache Software Foundation (ASF) announced Apache Ozone 2.0.0, a solution for the cloud-native distributed object store built for big data, analytics, and AI workloads. Key highlights and features of the new release, Apache Ozone 2.0.0, include:

Operational Database Support: Apache Ozone supports Apache HBase, enabling low-latency read/write workloads alongside traditional object storage.

Atomic Key Operations: Support for atomic key overwrite and key replacement improves consistency for applications running concurrently.

Modernized Recon UI: The Ozone Recon monitoring interface has been redesigned for clarity, with better metrics and navigation to support administrators.

Expanded Platform Support: Now compatible with JDK 17 and JDK 21, and builds natively on ARM64, enabling broader deployment options.

Improved Snapshots: Snapshot operations are more robust and efficient, particularly in large-scale environments with frequent replication or churn.

Alation introduced Alation Chat with Your Data, enabling anyone to get fast, accurate answers from structured data by asking questions in plain English. Unlike standalone solutions, Alation’s metadata-aware agents use business definitions, context, and lineage to reason through each response, improving accuracy. Each response also includes a clear explanation of how it was generated.

Appian announced enhancements that help organizations work smarter with faster insights, greater scalability, and more secure AI access. Key updates include AI-powered semantic smart search, Appian AI availability for self-managed and FedRAMP environments, automatic data fabric scaling, and Process HQ reports that embed directly into sites.

Composio announced its Universal MCP Gateway. The enterprise-grade platform solves the security challenges preventing organizations from deploying AI agents at scale by replacing hundreds of third-party MCP servers with a single, fully-implemented gateway. Specifically, Composio’s MCP Gateway Platform addresses many pain points by providing a secure, managed solution that eliminates risks from unvetted servers while simplifying deployment and management.

Confluent announced Streaming Agents, a new capability in Confluent Cloud for Apache Flink that makes it easy to build and scale AI agents that monitor, reason, and act on real-time data. Streaming Agents removes barriers to enterprise-grade agentic artificial intelligence (AI) by unifying data processing and AI workflows and providing easy, secure connections to every part of a business, including large language models (LLMs) and embedding models, tools, and other systems.

Druva announced an expansion to DruAI, the company’s suite of AI capabilities for customers. Built with Amazon Bedrock AgentCore on AWS, DruAI features DruAI Agents, intelligent agents that can interpret user intent, analyze data, and take meaningful action. This shift moves enterprises beyond traditional, query-based AI to agentic systems designed for action.

FPT announced the launch of its new AI platform, FleziPT. The solution empowers organizations to achieve exceptional speed, precision, and quality as they optimize for growth and expansion. The effort (and others by the company) are bolstered by its strategic partnerships with leading global AI players, including Microsoft, SAP, Landing AI, and more.

LambdaTest launched the private beta release of its Agent-to-Agent Testing, a platform designed to validate and assess AI agents. With the rise of AI agents in developer workflows, the platform helps organizations test and validate their AI agents across conversation flows, intent recognition, tone consistency, complex reasoning, and beyond. The platform highlights key metrics like Bias, Completeness, Hallucinations, etc., to help teams analyze the quality of their AI agents.

Oracle has deployed OpenAI GPT-5 across its database portfolio and suite of SaaS applications, including Oracle Fusion Cloud Applications, Oracle NetSuite, and Oracle Industry Applications, such as Oracle Health. By uniting trusted business data with frontier AI, Oracle is enabling customers to natively leverage sophisticated coding and reasoning capabilities in their business-critical workflows.

pgEdge announced the availability of pgEdge Platform v25. The solution builds upon Platform v24 advancements in Postgres logical replication. New features in this release include true zero downtime for node addition and PostgreSQL upgrades, expanded automatic conflict resolution, improved performance, interactive installation that guides customers through the needed information and creates the configuration file for them, and more.

SurrealDB announced the launch of SurrealMCP, a Model Context Protocol (MCP) server for SurrealDB and SurrealDB Cloud. SurrealMCP gives AI assistants, AI agents, IDEs, chatbots, and data platforms the ability to securely store, recall, and reason over live structured data, giving them the persistent, permission-aware memory.

Tredence announced the launch of Milky Way, a multi-agent,multi-turn agentic decision system that transforms enterprise decision-making using autonomous AI agents. Milky Way combines Tredence’s decade of domain expertise with a robust architecture featuring 15+ prebuilt agents tailored across critical business roles and 50+ specialized agents all trained on real-world enterprise scenarios.

VAST Data announced VAST SyncEngine, a new capability of the VAST AI OS that acts as a universal data router, unifying a high-performance onboarding solution with a global catalog to accelerate the flow of data into AI pipelines. Offered at no additional cost to VAST customers, SyncEngine eliminates the friction of discovering and mobilizing distributed unstructured datasets and enterprise SaaS platforms, so organizations can move faster from raw data to real-world AI outcomes.

Partnerships, collaborations, and more

dbt Labs announced the launch of its newly architected global partner ecosystem program. It introduces structured partner tiers, unified go-to-market models, and increased investment in enablement, laying the foundation for predictable, scalable growth across the dbt Labs partner ecosystem.

Deepgram announced that it has signed a strategic collaboration agreement (SCA) with (AWS). As part of the collaboration, Deepgram will expand co-selling and go-to-market efforts, integrate more deeply with AWS services, and empower enterprises to build scalable, high-accuracy voice applications across a wide range of use cases.

Hitachi Vantara announced the availability of Virtual Storage Platform One Software-Defined Storage (VSP One SDS) in the Microsoft Azure Marketplace, an online store that provides applications and services for use on Azure.

Precisely announced a new strategic technology partnership with Opendatasoft. Together, they will deliver an integrated data marketplace designed to simplify access to trusted, AI-ready data across businesses and teams, seamlessly and in compliance with governance requirements.

If your company has real-time analytics news, send your announcements to [email protected].

In case you missed it, here are our most recent weekly real-time analytics news roundups: