BentoML Raises $9M in Seed Funding to Accelerate AI Application Development

The rise of large language models like GPT has created a demand for AI-powered applications and a need for skilled developers to create them. Former Databricks software engineer Chaoyu Yang noticed the gaps in the current development process and, alongside his co-founders, built the AI development framework, BentoML, closing a seed financing round recently.

In an interview with TechCrunch, Yang explained the need for the flexible development framework BentoML provides. AI services tend to rely on multiple machine learning models, complicating their management and operation. Moreover, many programmers entering the field come from full-stack or application development backgrounds, lacking the skills required to build the necessary AI infrastructure. This causes longer development processes, and a solution like BentoML can ease the development process significantly.

Yang used Microsoft's Visual ChatGPT, a demo AI app that seamlessly combines text and image prompts to generate responses, as an example. Developing an app like Visual ChatGPT for production use generally takes three to six months. However, BentoML can enable developers to make an app like this scalable and cost-efficient in as little as two days. Users have also employed the BentoML framework to run the art generator Stable Diffusion and open-source LLMs on the cloud.

BentoML, which offers a high-level API that abstracts the intricate infrastructure needed for running AI models on the cloud, is part of a growing group of tools aiming to simplify AI service development. These AI application frameworks, which include solutions like AWS's SageMaker and AppMaster, provide a set of tools that ease the process of building, shipping, and scaling AI applications. The target audience for BentoML includes data scientists training AI models, DevOps engineers managing their lifecycle, and developers building applications upon these models.

Comparing BentoML to Vercel, a billion-dollar company focusing on front-end developers, Yang aims for BentoML to become the Vercel of the AI world. He recognizes the speed at which AI application development has grown and the increasing reliance on model fine-tuning and product engineering rather than training models from scratch.

Initially open-sourced in 2019, BentoML introduced a self-hosted SaaS version to enterprise customers. The startup has been gaining traction in the developer community, quadrupling its open-source community membership to over 3,000 within the last year. Early adopters of BentoML include South Korean social networking giants Line and Naver.

BentoML's recent seed funding round, backed by DCM Ventures and Bow Capital, raised $9 million. Hurst Lin, DCM's general partner, joined BentoML's board after the round. The growing AI market presents both opportunities and challenges for BentoML as the team works to balance short- and long-term goals while staying ahead of the evolving industry trends.