The startup offers a Python-centric feature store that enables data teams to efficiently build and manage feature, training, and inference pipelines at scale. This platform enhances the performance and availability of AI and machine learning applications across diverse data sources and environments, facilitating the development of data-driven products for business growth.
Funding
$13.9M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Data science teams face challenges in efficiently building, managing, and scaling feature pipelines for AI and machine learning applications. Existing solutions often lack the necessary tools for feature engineering, storage, and real-time serving, leading to increased development time and operational complexity.
Solution
Hopsworks provides an AI lakehouse platform centered around a feature store, enabling data teams to streamline the development and deployment of reliable AI systems. The platform offers a collaborative environment for feature engineering in Python, with support for various frameworks like Spark and Flink. It centralizes feature management, governance, and versioning, ensuring data consistency and reusability across projects. Hopsworks facilitates real-time feature serving with sub-millisecond latency and provides end-to-end MLOps capabilities for orchestrating, monitoring, and managing machine learning workflows.
Target Audience
The primary audience includes data scientists, machine learning engineers, and data engineers who need a comprehensive platform for building, managing, and deploying feature pipelines and AI models at scale.
Features
- Python-centric feature engineering environment with support for Pandas, Scikit-Learn, TensorFlow, and PyTorch
- Unified feature store for centralizing, sharing, and reusing ML features at scale
- Real-time feature serving with sub-millisecond latency powered by RonDB
- End-to-end MLOps capabilities for model training, registry, and serving via KServe
- Project-based multi-tenancy for secure collaboration and sharing of ML assets
- Support for feature pipelines in PySpark, Spark, Flink, and SQL
- Integration with external data lakehouses like Snowflake, Databricks, and Redshift via External Feature Groups
- Vector database based on OpenSearch for similarity search capabilities
- Integrated logging and monitoring for drift detection, performance metrics, and alerting
- GPU management for maximizing GPU utilization for LLMs and deep learning