Materialize’s cover photo
Materialize

Materialize

Software Development

New York, NY 8,539 followers

Transform siloed data into up-to-the-second context, just using SQL

About us

Materialize is the live context layer for agents and apps. It lets engineering teams use SQL to transform siloed operational data into real-time context they can trust. Organizations from Notion to Crane Worldwide Logistics use Materialize to deliver fresh context for AI agents, streamline search pipelines, and offload complex queries from OLTP databases.

Website
https://materialize.com
Industry
Software Development
Company size
51-200 employees
Headquarters
New York, NY
Type
Privately Held
Founded
2019
Specialties
Data Intensive UIs, Event-Driven Architecture, Digital Twins, AI Search Pipelines, Context Engineering, Loop Engineering, and Microservice Architectures

Products

Locations

Employees at Materialize

Updates

  • View organization page for Materialize

    8,539 followers

    How do you turn siloed operational data into trustworthy context for interactive agents? That's one of the big questions we set out to answer with Bilt. Their live context layer is now in production, behind the agents and microservices that power their commerce network. Learn how an up-to-the-second context graph keeps RAG and vector search fresh, and lets their engineers ship in days what used to take months, on smaller and cheaper models. Read the full post: https://lnkd.in/gemj5YPE

    How do you turn siloed operational data into trustworthy context for interactive agents? What would you encode in the data model itself to get more out of every token? These are some of the big questions we set out to answer as we partnered with Bilt to create a live context layer for the agents and microservices that make up their commerce network. It’s running in production and unlocking new experiences for millions of members along with new ways for their merchant partners to engage customers in real time. The architecture: - keeps RAG and vector search continuously fresh, incrementally updating only the documents affected by each human or agent action - maintains an up-to-the-second context graph that long-running agents can reason and act over - improves token efficiency at scale, letting them optimize their inference economics by blending in more cost-effective models Bilt’s engineers now ship faster: projects that once took months ship in days. Each new context building block expands their agents’ worldview and makes the next use case easier to build, creating a flywheel for new capability development. Here’s a deep dive into how they did it: https://lnkd.in/gemj5YPE

    • No alternative text description for this image
  • Great conference at Ai4 so far! Yesterday, our Field CTO Seth Wiesman gave a talk on building interactive vector pipelines just using SQL, to a packed room. Today, we've had some great conversations on powering agents with fresh, trustworthy context, and we're also hosting a Happy Hour from 6–8 PM. Come and find us at booth #1055 to learn more!

    • No alternative text description for this image
  • Materialize reposted this

    Ai4 - Artificial Intelligence Conferences takes over Las Vegas this week. Who’s going to be there?? Materialize will be at Booth 1055. On Tuesday from 12:20–12:40 PM, our Field CTO, Seth Wiesman, will be speaking in the AI Agents (Technical) track. His session, “Powering Agents with Search: Creating Interactive Vector Pipelines with SQL,” will cover how SQL-based pipelines can keep search indexes and vector databases continuously fresh for #AI agents. We’re also hosting a Materialize Happy Hour on Wednesday from 6–8 PM at Sala 118 in The Venetian. If you’re attending, come meet the team and sign up for the happy hour here: https://lnkd.in/gEBjUFus

    • No alternative text description for this image
  • Our July releases make it easier to run context engineering workloads reliably at scale and faster for coding agents to build on Materialize. Highlights include: ⚡ New autoscaling that spins burst replicas up and down to accelerate hydration 🤖 New query capabilities for our MCP servers and browser-based OAuth, making it easier for coding agents to build and debug against real data 📈 13% more queries per second under heavy concurrent write load and twice as many objects in self-managed deployments 🧊 Iceberg sinks on Google Cloud and AWS Glue Schema Registry support for Kafka sources and sinks 🐘 Support for replicating from PostgreSQL hot standbys We’ll also be at Ai4 - Artificial Intelligence Conferences in Las Vegas this month. Join our Field CTO, Seth Wiesman, for his talk on building interactive vector pipelines using only SQL. You can find us at booth #1055. Read the full July release notes: https://lnkd.in/gKwV6eUE

  • View organization page for Materialize

    8,539 followers

    Where are your AI agents actually spending their time: reasoning and executing tasks, or getting the right context before they can even begin? Our new post breaks down the hidden costs of agents discovering, joining, and transforming data, and how teams are avoiding them in production: the live context graph. Learn how a graph of real-time data products lets your agents discover and assemble context quickly and efficiently, and run on smaller, cheaper models as a result. Read the full post: https://lnkd.in/eQSwtBHk

  • Where should the work of turning raw data into agent context happen: at read time, at write time, or somewhere in between? Our new post breaks down what each option costs, in tokens, staleness, or pipeline complexity, and the pattern emerging in production to escape all three: the live context graph. Learn how to give your agents the fresh context they need, just using SQL, via a graph of interconnected, real-time data products. Read the full post: https://lnkd.in/eWQuWaD8

  • If you're wrestling with stale, rigid, or expensive search pipelines, this one's worth a watch.

    I've been having more conversations with engineering teams to help them streamline their search and vector pipelines to tighten agent loops and deliver the freshness required to support interactive RAG. It turns out that handling agent-scale writes while also delivering the right embeddings and attributes is a genuinely hard problem. I packaged up the pattern we've been rolling out into a short video.

  • How do you keep search indexes and vector databases fresh enough for an agent to act on? In his new blog post, Materialize's Seth Wiesman talks through what changes when agents become the primary consumer of search, and how to avoid the impact of stale results or complex pipelines. Materialize keeps your search documents up to the second using the SQL you already know. No batch processes, no streaming code to maintain. Read the full post: https://lnkd.in/gE84UJRt

  • We just published our latest newsletter. Here's what's inside: AI agents need a tight feedback loop. The teams building reliable agents in production aren't patching together pipelines. They're converging on a new pattern: the live context graph. We break down what it is, why it works, and how companies in SaaS, fintech, and logistics are running it today with Materialize. Plus what we've shipped lately: MCP servers for agents and developers, mz-deploy for declarative deployments, SSO on Self-Managed, and a batch of performance wins (75% CPU reduction on temporal filters, 65% faster DDL, and more). Read the full newsletter:

  • Agents need a tight feedback loop: observe fresh data → act → confirm → repeat. Traditional databases weren't built for this. 28 weekly releases later, here's what's new in Materialize for the agent era: - MCP Server for Agents: discover & query data products with per-agent RBAC - 2x higher QPS, 50% lower p99 latency - Replacement materialized views, iterate without downtime - MZ-deploy: declarative deployments, coding-agent friendly - Iceberg sink: exactly-once delivery to your lakehouse The live context graph your agents need. Read more about latest product releases: https://lnkd.in/g8Mt-t2D

    • No alternative text description for this image

Similar pages

Browse jobs