Tower is a Python‑native platform that unifies orchestration of data pipelines, dbt workflows, notebooks, and AI agents in a single managed service. It provides fully‑managed Apache Iceberg lakehouse storage compatible with Snowflake, Spark and other engines, along with built‑in observability, logs, alerts and metrics. By integrating popular ETL libraries and AI toolkits, Tower lets data and ML teams build, run, and monitor data products at scale with minimal infrastructure overhead.
Funding
Funding not disclosed
Founders
Product
Problem
Organizations building data products often struggle with fragmented tooling for pipeline orchestration, data lake management, and AI integration, leading to high operational overhead and limited scalability.
Solution
Tower offers a Python-native platform that unifies orchestration of data pipelines, dbt workflows, notebooks, and AI agents within a single service. Users can define control flows directly in Python or leverage agentic workflows, while the platform provides full observability through logs, alerts, and metrics. Tower also supplies a managed, open lakehouse built on Apache Iceberg, compatible with Snowflake, Spark, and emerging compute engines, handling table maintenance, storage optimization, and both batch and streaming ingestion. The service integrates popular ETL libraries (e.g., dltHub, Polars) and AI toolkits (OpenAI, LangChain, HuggingFace), enabling seamless data transformation and inference without additional infrastructure. By consolidating these capabilities, Tower reduces the effort required to develop, deploy, and maintain data applications at scale.
Target Audience
Primary customers are data engineering and analytics teams, as well as machine‑learning engineers, who need a unified, scalable environment to build, run, and monitor data pipelines and AI‑driven data products.
Features
- Pythonic orchestration API for defining pipelines, dbt commands, and agentic workflows in native code
- Managed Apache Iceberg lakehouse with REST catalog, automated table maintenance, and support for batch and streaming ingestion
- Built‑in observability suite offering real‑time logs, alerts, and performance metrics for all data flows
- Compatibility with major analytics engines (Snowflake, Spark) and open‑source ETL tools such as dltHub and Polars
- Integrated AI agent support, including connectors to OpenAI, LangChain, Ollama, llama.cpp, and HuggingFace for inference within pipelines
- Secure, cloud‑hosted execution environment that isolates compute and storage while providing role‑based access controls