Platform based on years of research performed at University of California San Diego
CTGT, enabling enterprises to finally scale their AI efforts with a new approach to customizing, training and deploying AI models that’s up to 500x faster, announced today that it has raised an oversubscribed $7.2M seed round to accelerate product development and sales & marketing.
Also Read: The Arbitrage Opportunity of Small Language Models: Unlocking AI Efficiency and Performance
Gradient, Google’s early-stage AI fund, led the round with participation from General Catalyst, Y Combinator, Liquid 2, Deepwater and notable angels including François Chollet (Google, creator of Keras), Michael Seibel (Y Combinator, co-founder Twitch), Paul Graham (Y Combinator), Peter Wang (co-founder Anaconda), Wes McKinney (creator of Pandas), Mike Knoop (co-founder Zapier), Kulveer Tagger (Zeus Living), Andrew Miklas (co-founder PagerDuty) and Taner Halicioglu (first full-time Facebook employee). This is the company’s first funding round.
As enterprises seek to move their AI projects from proof-of-concept to production at scale and move from low-risk use cases like chatbots to high-risk ones like security, the limits of AI compute have become apparent. AI requires enormous (and growing) amounts of compute and energy. Model developers speak of AI hitting the wall on compute, limiting what AI can do.
CTGT co-founder & CEO Cyril Gorlla has been pondering this challenge for years and made it the focus of his research for his endowed chair at the University of California at San Diego. In 2023, he published a seminal paper on the topic, presented at ICLR, that described a new way of evaluating and training AI models that was up to 500x faster and resulted in three nines of accuracy – a huge leap over current methods. That methodology became the basis for CTGT.
“CTGT’s launch is timely as the industry struggles with how to scale AI within the current confines of computing limits,” said Darian Shirazi, Managing Partner at Gradient. “CTGT removes those limits, enabling companies to rapidly scale their AI deployments and run advanced AI models on devices like smartphones. This technology is critical to the success of high-stakes AI deployments at large enterprises.”
While many other vendors can identify model problems, only CTGT can automatically refine and retrain models on the fly in production environments – eliminating the need to take models offline for updates.
Enterprises can use CTGT to ensure that AI models perform in line with their policies, including privacy, security and corporate standards guidelines – even as environments change. CTGT can help companies respond to changes in customer demand by giving models more autonomy or being more restrictive to security issues when new threats emerge.
Also Read: Combatting the rise in AI-assisted fraud in 2025
For instance, if an enterprise faced an emerging online security threat such as a prompt engineering attack, CTGT could recognize that and adjust a model on the fly to resist the attack. CTGT can also detect and fix hallucinations, inaccuracy and data leakage.
CTGT’s Gorlla said, “The lack of certainty and trust in models’ output is a significant barrier to adoption in high-stakes industries like healthcare and finance, where AI can make the biggest difference. By greatly improving accuracy, CTGT is removing that barrier.”
Founded in mid-2024 by Gorlla and co-founder Trevor Tuttle, CTGT is already working with a Fortune 10 company to deploy safe, on-device AI, and has landed enterprise customers who are already relying on CTGT software to close the gaps between AI safety and deployment.
One of CTGT’s first customers is Ebrada Financial, which leveraged CTGT to improve factual accuracy of its frontline customer service chatbots. “Previously, hallucinations and other errors in chatbot responses drove a high volume of requests for live support agents as customers sought to clarify responses,” said Ley Ebrada, Founder and Tax Strategist at Ebrada Financial. “CTGT has helped improve chatbot accuracy tremendously, eliminating most of those agent requests. We’re very happy with the performance.”
[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]
