Transforming Workplace Learning with AI and Neuroscience
The Challenge: L&D Content Creation, MLOps Capabilities and ChatGPT Introduction
A major challenge L&D platforms are facing is the efficient creation of learning and development content. Traditional AI technologies would be at the core of Stellar Labs L&D platform to automate content creation, enhancing personalization, and improving scalability.
For the AI Platform's infrastructure, there was a strategic aspiration to develop a custom-built MLOps platform to enhance the efficiency and effectiveness of content creation. The need for an MLOps platform arises from the complex nature of managing and operationalizing machine learning models, especially in the realm of AI-generated content. MLOps, or Machine Learning Operations, bridges the gap between the development of machine learning models and their deployment and maintenance in production environments. It facilitates continuous integration, delivery, and monitoring of ML systems, ensuring that AI-generated content is not only produced with the highest accuracy and relevance but also remains dynamic and adaptable to changing data and user needs. By leveraging MLOps, AI platforms can automate and streamline their content creation processes, reduce time-to-market for new features, and maintain high-quality standards, thereby significantly enhancing the scalability and reliability of AI-generated content.
As the Stellar Labs organization had yet to internalize AI and MLOps capabilities in its early stage, guidance was sought and a partnership with Faktion was formed. Both teams were set to go, the roadmap of the joint team had been developed, the funding secured and the governance was in place.
But then, the big plot twist: the launch of ChatGPT with out-of-the-box learning content readily available. Evidently, this sparked discussions about harnessing this new technology, about buy versus build, about pivoting.
Seizing the Opportunity: Leveraging LLMs and Generative AI
We quickly recognized the potential of this new generation of Large Language Models (LLMs) and Generative AI in transforming the L&D landscape. The impactful desision was made to pivot away from the original plan to build the MLOps platform and instead using the available LLMs and infrastructure as the new baseline for the MVP and first releases. So with all guns in place, a pivot was quickly decided and organized, enbling Stellar Labs to become a pioneer in using fully productized generative AI in the L&D space.
A seamless integration of LLMs into the Stellar Labs learning environment. The solution involved close collaboration between the two teams to create an LLM-based AI component within the platform, capable of generating customized learning journeys and evaluating the quality and consistency of the L&D content.
The Approach: A Four-Step Journey to Success
Following Faktion's four-step methodology, the initial phase involved comprehensive research on Large Language Models (LLMs) and generative AI within the learning and development (L&D) sector. The teams from Faktion and SL thoroughly explored how such technology could be integrated smoothly into the platform, how to optimize prompt creation, and how to set up continuous validation processes.
After completing the scoping and analysis phase, a Proof of Concept (PoC) was developed. This PoC addressed a subset of the full scope's challenges, utilizing a portion of the available data. This PoC was then expanded into a Proof of Value (PoV), encompassing the entire scope but not yet fully operationalized. A successful PoV laid the groundwork for deployment, applying innovative LLMOps principles to evolve the application into a fully developed solution.
The Result: Stellar Labs as a first-mover with GenAI in L&D
The outcome of the first development phase was a fully operational LLM-based learning journey generation model, implemented in record time. Stellar Labs internalized the necessary AI & MLOps capabilities as well, securing a valuable first-mover advantage in the industry. And the L&D professionals could generate personalized learning journeys efficiently and cost-effectively, while ensuring high-quality content. L&D journey content creation was being reduced from weeks to hours.