As early as fall 2023, driven by the need to meet real-world production demands while upholding open-source integrity, Oxide AI set out to transition from proprietary LLMs like OpenAI to a more scalable, transparent and efficient AI ecosystem, capable of delivering SOTA Gen AI at scale.
By leveraging IBM Cloud and watsonx.ai, Oxide AI successfully integrated open-source Llama models into live publishing flow of Oxogen application, achieving a 37% faster response time and a 95% qualitative acceptability threshold.
All within a month!
This shift enhanced control, increased stability, reduced the carbon footprint and laid the foundation for a deeper research partnership, positioning Oxide AI to gain a competitive edge in financial data intelligence through fine-tuning specialised language models.
of open, enterprise-ready generative AI alternative
with optimized model implementation
with enhanced control and transparency
This proof-of-value deployment marked just the beginning of Oxide AI’s journey in leveraging the latest generative AI solutions in scalable, enterprise-grade setups. Building on this success, Oxide AI has upgraded its proprietary platform, Polychaos®, to run IBM Cloud as well as Amazon Bedrock in AWS cloud, with further generalization of being cloud agnostic.