This Startup is Convinced That Compact, In House AI Models Will Prevail

ZenML aims to serve as the adhesive that connects various open-source AI tools. This open-source framework enables data scientists, machine-learning engineers, and platform engineers to collaborate and construct new AI models using shared pipelines.

This Startup is Convinced That Compact, In House AI Models Will Prevail
This Startup is Convinced That Compact, In House AI Models Will Prevail

This Startup is Convinced That Compact, In House AI Models Will Prevail

ZenML stands out because it enables companies to create their own exclusive models. While they might not aim to develop a GPT-4 competitor, they can construct smaller models tailored to their specific requirements. This approach reduces their reliance on API providers like OpenAI and Anthropic.

Louis Coppey, a partner at VC firm Point Nine, explained, “The concept is that, once the initial wave of excitement with everyone using OpenAI or closed-source APIs subsides, [ZenML] will allow individuals to construct their own technology stack.

Earlier this year, ZenML extended its seed round funding with an investment from Point Nine, with existing investor Crane also participating. In total, the Munich-based startup has raised $6.4 million since its inception.

The founders of ZenML, Adam Probst and Hamza Tahir, had previously collaborated on a company that developed ML pipelines for a specific industry. “Day in, day out, we needed to build machine learning models and deploy them,” explained ZenML CEO Adam Probst.

Inspired by their experience, the duo began creating a versatile system that could adjust to various situations, settings, and clients, eliminating the need for repetitive tasks. This initiative gave rise to ZenML.

Simultaneously, novice machine learning engineers could benefit from this adaptable system to accelerate their work. The ZenML team labels this realm as MLOps, akin to DevOps but tailored for machine learning.

We’re uniting open-source tools that focus on specific value chain steps to construct a machine learning pipeline. This runs on the infrastructure of hyperscalers like AWS and Google, as well as on-premises solutions,” Probst explained.

ZenML: Revolutionizing Machine Learning Workflows

The central concept of ZenML revolves around pipelines. After creating a pipeline, you can execute it locally or deploy it using open-source tools like Airflow or Kubeflow. Additionally, you can leverage managed cloud services such as EC2, Vertex Pipelines, and SageMaker. ZenML also integrates with various open-source ML tools from Hugging Face, MLflow, TensorFlow, PyTorch, and more.

ZenML serves as the unifying element, offering a comprehensive, multi-vendor, and multi-cloud experience, as noted by ZenML CTO Hamza Tahir. It enhances ML workflows by introducing connectors, observability, and auditability.

Initially, the company introduced its framework on GitHub as an open-source tool, amassing over 3,000 stars on the platform. ZenML has also recently introduced a cloud version with managed servers, and there are plans to add triggers for continuous integration and deployment (CI/CD).

Several companies have employed ZenML for various use cases, including industrial applications, e-commerce recommendation systems, and medical image recognition. Notable clients include Rivian, Playtika, and Leroy Merlin.

Check These Out


Please enter your comment!
Please enter your name here