TrueFoundry, a startup established by former Meta engineers, is focused on assisting enterprises in deploying AI systems at scale and has recently secured $19 million in new funding, primarily from Intel Capital.
Since the introduction of ChatGPT in November 2022, the rise of Generative AI has prompted businesses of all sizes to explore how they can leverage AI technologies. Nevertheless, Generative AI requires substantial computing power due to its dependency on large language models, which not all companies can afford, particularly those lacking access to extensive GPU resources. Consequently, optimizing GPU use becomes vital. Furthermore, enterprises must experiment with various models and refine existing ones to ensure their AI systems effectively integrate with their operational needs.
TrueFoundry tackles these challenges by offering its software as a platform-as-a-service, specifically designed for full-stack data scientists. Key features include auto-scaling, proactive maintenance, centralized access management, and real-time monitoring, all aimed at streamlining comprehensive AI deployments.
Founded in June 2021 by former Meta engineers Nikunj Bajaj and Abhishek Choudhary alongside their IIT Kharagpur peer Anuraag Gutgutia, TrueFoundry initially developed a cross-cloud native software to accelerate machine learning deployment. However, with the advent of Generative AI in 2023, the startup adapted its platform to support GenAI functionalities.
“Previously, data scientists focused primarily on modeling and experimentation, while much of the deployment work was relegated to ML engineers. Our system empowers data scientists to construct and test complex systems at scale before collaborating with the platform team for final deployments,” Gutgutia shared in an exclusive interview.
The startup’s software solution, an autopilot system, leverages AI to analyze logs and metrics, optimizing AI application scalability by dynamically adjusting GPU allocations and meeting memory needs.
Through this autopilot system, one of TrueFoundry’s clients, a small team of just two people, successfully handled 10 million requests per second, as highlighted by Gutgutia during his conversation with TechCrunch.
Currently, TrueFoundry serves 30 paying clients globally and has many users experimenting with its open-source RAG framework. Among its notable customers is NVIDIA, which utilizes their software to construct and deploy agents that optimize GPU clusters. The company Resmed, a player in the med-tech field, is also utilizing TrueFoundry to build its internal AI platform. Other clients include Siemens Healthineers, Automation Anywhere, Games24x7, and Whatfix.
The startup has engineered its software using Kubernetes, ensuring it is multi-cloud ready and compatible with leading platforms, including AWS, Google Cloud Platform, and Azure. Additionally, enterprises have the option to deploy the platform natively on-premises.
Enterprises leveraging TrueFoundry are able to develop and launch their internal AI platforms within two months and see their return on investment in just four months—significantly faster than the industry standard of 14 months, according to Bajaj. Gutgutia noted that the startup has helped its customers achieve a 40-50% reduction in infrastructure costs and enhanced deployment speed by 10 times, whether they are focusing on an agent, RAG, complex deep learning models, or fine-tuned models.
While cloud service providers have begun offering their own tools for simplifying the construction and deployment of AI models, such as Amazon’s Sagemaker and Google Cloud’s Vertex, Gutgutia emphasized that TrueFoundry remains relevant as it aids in “accelerating compute usage” for these cloud providers, hastening the time-to-market for AI applications, and minimizing overall costs.
The recent Series A funding round, which was entirely equity-based, saw participation from Eniac Ventures, Peak XV Partners, and Jump Capital, alongside angel investors like Gokul Rajaram, Mohit Aron, and Cyan Bainster.
This funding will enable the startup to grow its team from 45 members in India to include sales, customer success, and product marketing roles in the U.S. TrueFoundry also plans to strengthen partnerships, particularly with cloud vendors, and is working towards being featured in major cloud marketplaces, starting with AWS Marketplace.
Furthermore, TrueFoundry intends to roll out its AI agent, which will assist in resource recommendations, enable auto-scaling, and manage troubleshooting.
With this latest funding, TrueFoundry’s total capital raised has exceeded $21 million, including the $2.3 million from its seed round in September 2022, which was led by Peak XV’s (previously Sequoia India) Surge program.
Last year, the startup rapidly grew its customer base, quadrupling the number of clients and deploying over 1,000 clusters for machine learning workloads. Gutgutia mentioned to TechCrunch that the company is currently above $1.5 million in annual recurring revenue and aims to double that figure this year.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
