Follow Datanami:
May 3, 2023

Appen Partners With Reka AI to Build Custom Enterprise LLM Applications

KIRKLAND, Wash., May 3, 2023 — Appen Limited, the global leader in data for the AI Lifecycle, today announces its partnership with Reka AI, a developer of generative AI large language models (LLMs), for an unrivaled combination of Appen’s world-class data services with Reka’s proprietary multimodal language models.

As a result of innovations like ChatGPT, the implementation of LLMs has skyrocketed. With LLMs, businesses can improve the productivity of their operations and deliver new and engaging experiences for their customers. However, large organizations encounter significant friction during implementation as they are not an out-of-the-box solution for enterprises. To leverage the power of LLMs, businesses need to fine-tune models for their use cases, and continuously evaluate and monitor the performance of these models in the real world.

With the combination of Appen’s and Reka AI’s expertise, enterprises can now create and own enterprise-grade production models that meet their specific deployment requirements. With 26 years of deep expertise in AI training data and language, Appen is uniquely positioned to accelerate the LLM journey for enterprises, enabling them to unlock the full potential of generative AI for their businesses. Coupled with Reka’s advanced proprietary algorithms, the two companies have developed the most effective and comprehensive generative AI solution for enterprises to date.

“Our partnership and technical integration with Reka provides leading enterprises with the ability to build custom proprietary models with high levels of security.” said Armughan Ahmad, CEO of Appen. “Enterprises typically rely on public APIs to access generative AI models, leaving them vulnerable to data leaks and privacy concerns for highly sensitive data. Appen and Reka provide enterprises the ability to operate LLM applications while meeting risk and assurance guidelines.”

Creating generative AI solutions requires deep expertise in data curation to leverage and supplement internal datasets, human feedback to improve performance, and a robust model evaluation platform. With Appen’s extensive experience with model evaluation, linguistic experience and access to a large and diverse global crowd, Reka is able to build, test and deploy its LLMs more quickly. Reka’s proprietary algorithms then enable rapid customization to many use cases. This partnership allows businesses to have a full stack solution to utilize Reka’s models for enterprise use cases.

“Our flexible approach allows businesses to deploy our models under different quality, latency, and privacy constraints,” said Dani Yogatama, CEO of Reka. “Partnering with Appen enables customers to further benefit from their world class expertise in data services. This will streamline the process to create production-ready applications.”

For more information about how the collaboration provides enterprises with proprietary ownership of the model and the data to meet their specific use case, visit

About Appen Limited

Appen is the global leader in data for the AI Lifecycle. With over 25 years of experience in data sourcing, data annotation, and model evaluation by humans, we enable organizations to launch the world’s most innovative artificial intelligence systems. Our expertise includes a global crowd of over 1 million skilled contractors who speak over 235 languages, in over 70,000 locations and 170 countries, and the industry’s most advanced AI-assisted data annotation platform. Our products and services give leaders in technology, automotive, financial services, retail, healthcare, and governments the confidence to launch world-class AI products. Founded in 1996, Appen has customers and offices globally.

About Reka

Reka is a full stack model provider that offers solutions to create enterprise-grade production models from their foundation models. Prior to Reka, their team was involved in some of the largest breakthroughs in AI research at DeepMind, FAIR, and Google Brain. Their multimodal large language models are trained to read text, images, and tabular data. Reka uses proprietary algorithms to rapidly and cost-efficiently customize these models to many data and use cases. Their versatile approach to personalization and compression allows clients to meet various deployment constraints—such as quality, latency, and data privacy requirements.

Source: Appen