DualBird offers a cloud‑native GPU acceleration layer that integrates directly with existing Apache Spark and Iceberg deployments without code changes. It automatically optimizes shuffle handling and prevents disk spills, delivering 10‑30× faster query execution while cutting cloud compute costs by 50‑90%. The platform includes a real‑time dashboard that reports cost and latency metrics for batch and AI workloads on major cloud providers.
Funding
$16.5M 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 pipelines built on Apache Spark and Iceberg often suffer from long runtimes, high cloud compute costs, and operational overhead caused by manual tuning, shuffle bottlenecks, and disk spills. These inefficiencies limit the ability of data teams to deliver timely insights and scale AI workloads cost‑effectively.
Solution
DualBird delivers a cloud‑native acceleration layer that plugs directly into existing Spark and Iceberg deployments, providing hardware‑level GPU acceleration without requiring code changes or cluster migration. The platform automatically optimizes shuffle handling, eliminates disk spills, and balances workloads to achieve 10‑30× faster query execution while cutting infrastructure spend by 50‑90%. Because integration is zero‑effort, data engineers can realize performance gains immediately through a few configuration steps, freeing them to focus on analytics rather than performance tuning. DualBird also offers a unified monitoring dashboard that surfaces cost and latency metrics in real time, enabling predictable scaling for both batch and AI workloads.
Target Audience
Primary customers are enterprise data engineering and analytics teams that run large‑scale Spark/Iceberg workloads on cloud platforms, as well as AI/ML pipelines that need fast, cost‑effective data preparation.
Features
- Drop‑in Spark/Iceberg plugin that leverages GPU‑accelerated execution engines with no code modifications required
- Automatic shuffle reduction and spill avoidance that removes the need for manual tuning of Spark configurations
- Cloud‑native deployment model compatible with Amazon EMR and other major cloud providers, supporting existing cluster sizes
- Real‑time performance and cost analytics dashboard with alerts for sub‑optimal workloads
- Built‑in cost optimizer that dynamically selects the most efficient compute resources, delivering 50‑90% lower EC2 spend
- Seamless API integration for CI/CD pipelines, enabling automated provisioning and scaling of accelerated jobs