The company's goal is to empower businesses to build their own private models, reducing their dependence on API providers like OpenAI and Anthropic. ZenML's founders believe that the majority of the market will need its own solution, and open source is very appealing to them. The company's clients include Rivian, Playtika, and Leroy Merlin, and it has been used for industrial use cases, e-commerce recommendation systems, and image recognition in a medical environment.
Key takeaways:
- ZenML is an open-source framework that allows companies to build their own AI models, reducing their dependence on API providers like OpenAI and Anthropic.
- The company has raised $6.4 million since its inception and is based in Munich, Germany. Its founders previously worked on building ML pipelines for other companies.
- ZenML's main concept is pipelines, which can be written and then run locally or deployed using open-source tools. It integrates with open-source ML tools from Hugging Face, MLflow, TensorFlow, PyTorch, and others.
- Despite the sophistication and cost of APIs like OpenAI, ZenML believes that the majority of the market will need its own solution, and that open source is very appealing for this reason. The company also believes that 99% of AI use cases will be driven by more specialized, cheaper, smaller models that will be trained in-house.