deepset's journey began in 2018, a pivotal year when the AI world was buzzing with the transformative potential of models like Google's BERT and the foundational 'Attention is All You Need' paper. From these early days, the founders, Milos Rusic, Malte Pietsch, and Timo Möller, recognized the immense opportunity in custom language AI. They embarked on a mission to build sophisticated AI systems tailored for industry leaders, quickly establishing themselves as pioneers in applying natural language processing (NLP) to real-world business challenges. Their initial focus was on helping enterprises implement NLP services by adapting powerful language models to their specific domains and datasets.
Driven by the practical experience gained from these early projects, deepset identified a critical need for a more accessible and developer-friendly way to build and deploy NLP applications. This realization led to the creation of Haystack in 2020, an open-source Python framework. Haystack rapidly gained traction, becoming an enterprise standard for building AI agents and applications leveraging large language models (LLMs). It provided the crucial building blocks for developers to orchestrate various components and create custom solutions. Building on the success of Haystack, deepset continued to innovate, launching the deepset AI Platform (formerly deepset Cloud). This platform further streamlined the development process, offering tools for data management, infrastructure, and templates to accelerate the journey from concept to production for LLM-powered applications. Today, deepset has evolved into a global company, recognized for making AI adoption faster, easier, and more impactful for organizations across diverse sectors, including finance, legal, media, and government. They continue to champion the power of custom AI, empowering teams to solve their most critical challenges with speed, trust, and control, all built upon a strong open-source foundation and a commitment to enterprise-grade solutions.