Redis on Monday introduced Feature Form, a set of features aimed at providing database customers with a managed environment for training and running AI and machine learning workloads.
Redis acquired Featureform for an undisclosed sum in October 2025, adding a framework for managing, defining, and orchestrating structured data signals, such as real-time sensor readings and user interactions with websites, that can be used to inform AI and machine learning tools.
The preview release of Feature Form by Redis represents a redesign, expansion, and integration of Featureform’s functionality into Redis’ broader database platform to help advance your machine learning efforts.
Specifically, Feature Form’s specific capabilities include integrated batch and streaming data pipelines, multi-tenancy that allows separate machine learning teams to work on a shared Redis instance, upgraded security features such as role-based access control, and a redesigned user interface.
According to IDC analyst Devin Pratt, this new capability is important for Redis users because Feature Form gives AI and machine learning teams the managed ability to deliver functionality across training and inference workloads, and across teams and environments, rather than relying on homegrown pipelines.
“Feature Form makes sense because it moves Redis beyond high-speed services and further into day-to-day management of features,” he said. “Enabling functionality to be operationalized across teams and environments is one of the key production ML challenges for large enterprises.”
Stephen Catanzano, an analyst at Omdia, a division of Informa TechTarget, similarly pointed to the value of Feature Forms.
“The ability to unify fragmented workflows and provide a managed system for defining, orchestrating, and delivering functionality across training and inference workflows addresses a critical need,” he said. “This allows users to maintain consistency that was previously difficult to achieve, reducing operational burden and improving model reliability.”
San Francisco-based Redis is a database specialist that started as an open source project in 2011. Competitors include database specialists such as Aerospike, MongoDB, and SingleStore, as well as hyperscale cloud providers with database capabilities such as AWS, Google Cloud, and Microsoft.
A better foundation for AI
At a time when many companies are struggling to move their AI and machine learning projects into production, Feature Form is a new set of features designed to help Redis customers better manage the models essential to AI development.
Feature forms make sense because they move Redis beyond fast service delivery and further into day-to-day management of functionality. Making functionality operational across teams and environments is one of the key production ML challenges for large enterprises.
Devin PrattIDC Analyst
According to Redis, building the underlying machine learning model is no longer the main issue hindering the development of AI and machine learning tools. Instead, you run into problems when you try to deploy the model across your team or environment after it’s built.
Model drift, where a model’s training data becomes stale, machine learning pipelines that break down in production due to changes in the underlying data or deployment in a new environment, and governance gaps are some of the roadblocks that enterprises encounter.
The Feature Form is designed to help Redis customers keep model training and model delivery in sync after the models are in production in various enterprise environments and deployed to power AI and predictive analytics efforts. Meanwhile, some of these customers struggled to bring machine learning capabilities into production, which led to Redis’ acquisition of Featureform and its subsequent introduction, according to Simba Khadder, founder and CEO of Featureform and now Redis’ AI product lead.
“Our largest customers have told us about the operational challenges of bringing ML capabilities into production, including real-time data pipelines, version control and lineage, and training and service consistency,” he said. “At the same time, we were seeing teams building their own solutions on top of Redis to address these issues.…The feedback confirmed that we needed to go further.”
Functional forms include:
Integrate batch files and streaming data pipelines to reduce the amount of work required for developers and engineers to customize machine learning model pipelines.
A workspace for organizations that use a single Redis instance with multi-tenancy. Allow teams to separate work.
Fine-grained job control gives teams greater visibility into data changes before they write data to other systems or invisible changes can inadvertently impact production systems.
Improving access controls and security measures.
The new deployment model aims to reduce complexity while enabling advanced patterns.
A redesigned user interface that supports all new workflows enabled by feature forms.
Atomic Directed Acrylic Graphs (DAGs) are visual representations of data models and their connections to each other, and are updated to make it easier to see the history of changes.
Pratt said the integrated streaming data and batch file pipelines are probably the most valuable of the new features because they alleviate some of the custom engineering work required for machine learning teams that would otherwise struggle to coordinate separate pipelines.
However, he noted that while Feature Form is beneficial for Redis users, its functionality is not unique among database vendors, with AWS, Databricks, Google Cloud, and Snowflake offering similar tools for machine learning workloads. Still, Redis has the potential to differentiate itself by introducing improved governance and orchestration capabilities to a database platform already known for its low-latency service delivery and real-time data workloads, Platt continued.
“Feature Form provides a reliable way for Redis to stand out by making the platform more complete for production ML,” said Pratt.
Catanzano highlighted the integration pipeline and atomic DAG updates as Feature Form’s most valuable features, but similarly noted that while other database vendors offer similar capabilities, Redis’ new feature set for AI and machine learning is unique.
“The Redis Feature Form differentiates itself by directly integrating with the Redis real-time data platform,” he said. “This combination of sub-millisecond performance and enterprise-grade feature management is unique and positions Redis as a leader in production ML environments.”
Looking to the future
With Feature Form, currently in preview, the primary focus of Redis’ product development will be on building database capabilities that will allow customers to access information relevant to the context of their agent AI development initiatives, Kader said.
Agents require relevant data to perform as intended. However, that data is often difficult to discover and provide.
“We focus on one central problem: making context available to agents,” Kader said. “Most companies don’t have an agent problem; what they have is a context problem. … We’re focused on the fundamentals of solving these problems.”
Specifically, Redis is building a context engine that integrates structured and unstructured data, along with memory, into a real-time context layer that agents can invoke.
To further improve machine learning capabilities, Catanzano recommended that Redis add support for more prebuilt models and integrate its database with popular third-party machine learning frameworks such as TensorFlow and PyTorch.
“Additionally, by focusing on industry-specific solutions, such as customized feature stores for healthcare and finance, we may be able to attract new customers while deepening the value for existing customers,” he said.
Meanwhile, Pratt suggested that Redis enhance observability features related to Feature Forms to help users better check whether features are fresh, stable, and performing as expected before they impact model performance.
“A powerful next step for Redis is deeper feature observability, which will give our customers more confidence as they scale ML in production,” he said.
Eric Avidon is a senior news writer at Informa TechTarget and a journalist with more than 30 years of experience. He is responsible for analysis and data management.
