TheStage AI steps into the limelight after raising $4.5M to automate model fine-tuning – News

TheStage AI Inc., a startup that’s optimizing artificial intelligence to run better at lower costs, has just announced a $4.5 million funding.

Today’s round was led by Mehreen Malik, the founder of ML Capital Investment, and saw participation from other angel investors, including Dominic Williams, founder of the DFINITY Foundation and Nick Davidov, founder of Davidovs VC. Atlantic Labs and AAL VC also took part in the round.

The startup wants to upend the AI model fine-tuning process by automating it with AI itself. It uses advanced mathematical algorithms to optimize each model according to the task in hand, and claims that its approach can reduce the time it takes from months to hours, while improving neural network performance by up to five-times.

Its flagship technology is known as ANNA, which stands for Automatic NNs Analyzer, a system that leverages AI and discrete math to automatically adjust PyTorch models using techniques such as quantization, pruning and sparsification. In essence, it trims the fat to make AI models leaner and more performant.

Using ANNA, developers can fine-tune their AI models to meet their exact requirements in terms of performance, size and latency. In addition, the company provides access to so-called “Elastic models,” or pre-fine-tuned open-source models, available in a range of sizes. Customers can select the most appropriate model based on the required quality, speed and cost.

TheStage AI’s main goal is to help companies reduce the cost of deploying and running AI applications. It cites a study by McKinsey Co. that shows how up to 70% of the costs of AI stem from the underlying graphics processing unit-based infrastructure that powers them, creating a huge barrier to entry that many startups cannot get through.

By optimizing AI models to run more efficiently, that barrier suddenly becomes much more surmountable. Within its Model Library, TheStage AI currently lists dozens of optimized models, including various fine-tuned versions of the popular image generation model Stable Diffusion. Customers can therefore strike exactly the right balance in terms of cost and performance. In addition, customers can also bring their own models and use ANNA to optimize them.

The company likens it to how YouTube provides users with the option to watch videos at different resolutions, so they can choose the most appropriate stream based on their internet connectivity speed.

The startup adds that customers have the freedom to choose their own hardware for their AI models, as it supports infrastructures ranging from cloud-based GPU clusters to on-premises environments and even smartphones

TheStage AI was founded by four friends – Kirill Solodskih, Azim Kurbanov, Ruslan Aydarkhanov and Max Petriev – who previously worked together at Huawei Technologies Co. Ltd. There, they helped to develop cutting-edge AI model compression and acceleration technologies that were later integrated with the AI-powered cameras on the Huawei P50 and P60 smartphones.

Now, they want to bring this know-how to a much wider audience through TheStage AI, and their offering has already been warmly received by a number of early adopters, including the AI image-generation platform Recraft Inc. and the AI tutor startup Pratika.ai Co. In its collaboration with Recraft, TheStage AI reckons it was able to double the performance of that company’s most powerful models while reducing processing times by 20%.

TheStage AI Chief Executive Solodskih said the company can help any AI developer bring their ideas to life. “We’ve created a service that allows AI engineers and developers to compress, package and deploy models to any device, as easily as copy and paste,” he said.

Malik said she believes there’s a great opportunity for TheStage AI to become an essential tool for AI developers.

“The most successful solutions always combine hardware with the right software,” she said. “I’m helping the team with business strategy, and we plan to demonstrate significant product growth in the next several months.”

Images: TheStage AI