TORONTO — Could 23, 2024 — Agnostiq right this moment introduced an replace to Covalent, its superior computing platform. With the discharge of latest capabilities referred to as Operate Serve, Covalent turns into an enterprise-grade resolution for AI growth and productionalization that gives serverless AI mannequin coaching, tuning and inference in a single platform.
Clients can now take any open-source or proprietary mannequin, fine-tune it, host it, and serve it with full infrastructure abstraction, all from a single python pocket book.
With the explosion of GenAI and LLMs, many startups and enterprises battle to effectively practice and deploy fashions because of the lack of instrument standardization in addition to the complexity of the required computing infrastructure. By offering full infrastructure abstraction and a unified framework for end-to-end AI growth, Covalent gives customers a clean canvas to create any AI software they envision, making the productionalization of AI accessible to startups and enterprises, permitting customers to rework ideas into ready-to-ship merchandise, and considerably lowering time to market.
“Agnostiq is fixing the toughest a part of AI software deployment for any enterprise or start-up, which is orchestrating the accelerated compute infrastructure and AI growth pipeline,” says Oktay Goktas, CEO of Agnostiq. “Covalent provides builders unprecedented pace and agility to construct, iterate, and deploy something that they’ll think about, from multi-agent, multi-modal AI functions, to digital twins for actual world simulation. What used to take months with dozens of best-in-class engineers can now be achieved in hours from a single Python pocket book.”
Key options and functionalities of Covalent embody:
- Finish-to-end growth: Covalent is a complete AI growth platform, permitting customers to handle the total lifecycle of productionalizing AI functions, from mannequin coaching to deployment.
- Infrastructure abstraction and orchestration: Covalent gives full infrastructure abstraction, eliminating the necessity for DevOps and AIOps sort infrastructure work, permitting builders to concentrate on code.
- Serverless structure: Covalent infrastructure is solely serverless, permitting dynamic useful resource allocation for the total utilization of costly and scarce compute assets, leading to important value financial savings.
- Multi-agent orchestration: Covalent consists of native workflow orchestration performance, enabling customers to construct production-grade AI brokers in a scalable and cost-effective approach. (See example here)
Since launching in January of 2022, Covalent has centered on offering customers with simplified and scalable entry to cloud-agnostic high-performance computing infrastructure, together with CPUs, GPUs, and application-specific compute modalities. With Operate Serve, Covalent builds on these capabilities and introduces important new use instances and functions into the platform, notably for AI mannequin coaching, fine-tuning and inference workflows. Different use instances of the Covalent platform embody large-scale simulations, digital twins, robotics, and any job that requires specialised or excessive efficiency computing assets comparable to GPUs, ASICs and quantum computer systems.
Be taught extra about Covalent for AI on this blog. Join right this moment and get began with as much as 30 hours of GPU compute credit.
