Artificial intelligence inference startup Simplismart, officially known as Verute Technologies Pvt Ltd., today announced that it has completed $7 million in funding to expand its infrastructure platform and help businesses to deploy AI models more easily.

The Series A round was led by Accel and saw participation from Shastra VC, Titan Capital and prominent angel investors such as Akshay Kothari, co-founder of Notion Inc.

Simplismart has created what it says is a “rapid inference engine” that allows companies to optimize the performance of AI model deployments. The startup says it wants to be seen as a critical enabler of the transition from AI to traditional business operations. To do this, it seeks to address a number of challenges that prevent business adoption of AI, such as the performance trade-offs that many businesses are forced to make.

In a blog post, Amritanshu Jain, co-founder and chief executive officer of Simplismart, says companies are increasingly interested in adopting AI, but are struggling to derive much value from it. Part of the problem is that it’s not easy for companies to deploy AI themselves. An alternative is to use third-party application programming interfaces, he explained, but these are expensive, inflexible and pose data security concerns.

“Every business has different inference needs, and one solution doesn’t fit all,” Jain said. “APIs are not designed to accommodate bursty workloads and cannot adjust performance as needed. Businesses need to control their trade-offs between cost and performance. This will be the main reason for a shift to open source models, as companies prefer smaller niche models trained on relevant datasets rather than large generalist models to justify ROI.

Jain says few companies want to “rent their AI,” but says many are forced to do so because owning AI isn’t easy. To deploy large language models internally, companies face significant hurdles related to scaling their infrastructure, building a continuous integration and deployment pipeline, accessing computing resources, model optimization and profitability.

Currently, most companies use one of two commercially available solutions for their AI, but both solutions have limitations. For example, MLOps platforms enable model orchestration and delivery, but they do not provide an optimized environment for production AI, meaning businesses face severe performance limitations.

The alternative is to use cloud generative AI platforms, or “GPU brokers,” which provide optimized APIs and performance, but have serious data privacy and cost concerns.

Simplismart’s inference engine is designed to give businesses a new option, providing a standardized language that software engineers can use when building generative AI applications. Its main advantage is that it reduces the time it takes for models to respond to queries. It cites benchmarks that demonstrate its ability to run the open source Llama 3.1 8B model at a throughput of >440 tokens per second. This represents an impressive step forward in speed, and it’s coupled with a comprehensive MLOps platform designed for on-premises AI deployments.

According to Jain, there is a significant market for what the startup offers. He cites data that shows nearly 90% of companies’ machine learning projects never make it to production.

“Adoption of generative AI lags far behind the pace of new developments,” the CEO said. “This is because businesses face four bottlenecks: lack of standardized workflows, high costs leading to low ROI, data privacy, and the need for system control and customization to avoid downtime and limitations of other services.”

Simplismart’s declarative language is similar to Terraform and helps software teams with tasks like fine-tuning, deploying, and monitoring large-scale generative AI models. The platform helps standardize all of these workflows, ensuring teams can optimize the performance of their models.

Simplismart was founded in 2022 by Jain alongside Devansh Ghatak, the company’s chief technology officer. While Jain’s experience lies in cloud infrastructure, primarily from his time at Oracle Corp., Ghatak’s area of ​​expertise is in search algorithms, which were honed during his time at Google LLC.

In just two years, with less than $1 million in capital, Simplismart managed to create a powerful MLOps platform for deploying models featuring a high-performance inference engine that, according to the founder, is the most fast in the world. With it, businesses can build, refine, deploy and then run their AI models on-premises at fast enough speeds, improving performance without the cost and security concerns.

Simplismart says it wants to help businesses deploy custom generative AI applications with full control. It sees itself as providing the granular Lego bricks that companies need to build their own inference and deployment environments, so they can do just that.

To date, Simplismart already has around thirty clients who generate a combined turnover of $1 million on an annual basis. With today’s round of funding, Jain believes the company can reach $5 million by the first quarter of next year.

The money from today’s round will be a big help to Simplismart, and it’s earmarked for product development, recruiting, and investing in its sales and marketing efforts.

Accel partner Anand Daniel said more companies have started to realize the benefits of deploying and customizing AI models on their own infrastructure, such as controlling performance, costs , data security, confidentiality, etc.

“What blew us away was how their small team had already started serving some of the fastest growing generative AI companies in production,” he said. “This reinforced our belief that Simplismart has a chance to win in the massive but extremely competitive global AI infrastructure market.”

Image: Simplism