[L-R] Instill AI Co-Founders Xiaofei Du (COO) and Ping-Lin Chang (CEO)
There are no simple tools in the market to tap into the value of unstructured data (images, videos, audio, and text data) easily.
“Building in-house AI solutions requires a tremendous investment of time and money and the intrinsic transformation of the team culture,” says Ping-Lin Chang. “Only Big Techs have such a luxury. This is where Instill AI comes into the picture.”
Instill AI provides no-code/low-code tools to convert unstructured data into meaningful data representations. Users can integrate its service into their data stack, tap into the wealth of their unstructured data, and benefit from AI in a snap.
The startup was founded in 2020 by Ping-Lin Chang and Xiaofei Du. Chang (CEO) holds a PhD in Robotic Vision from Imperial College London with a research focus on Visual Simultaneous Localisation and Mapping and Machine Learning for Augmented Reality in image-guided surgery. Meanwhile, Du holds a PhD in Medical Physics from University College London with a research focus on Surgical Vision and Medical Image Analysis with Machine Learning.
Last August, Taiwan-incorporated Instill AI launched its open-source project Versatile Data Pipeline (VDP). According to Chang, VDP is the future for unstructured data infrastructure, where developers won’t need to build their own data connectors, high-maintenance model serving platform, or data pipeline automation tool.
Also Read: Will AI replace humans in customer service?
“Our mission is to make VDP the single point of unstructured data integration to streamline the end-to-end unstructured data processing pipeline. VDP can extract unstructured data from pre-built data sources such as cloud/on-prem storage or IoT devices and transform it into analysable or meaningful data representations by AI models imported from various ML platforms. It can also load the transformed data into warehouses, applications, or other destinations,” Chang explains.
VDP currently supports popular AI tasks, including image classification, object detection, keypoint detection, optical character recognition and instance segmentation. More AI tasks, such as text generation and text-to-image, will soon be released.
“Our ultimate goal with VDP is to streamline the end-to-end unstructured data flow, with the transform component being able to import AI models from different sources flexibly,” Chang states.
Chang claims that Instill AI’s solution can be modularised into working components to benefit a broader spectrum of AI tasks and industry sectors.
“To be more specific, our no-/low-code unstructured data pipeline solution can significantly save development resources to harness the latest AI technology for AI-first application companies (who build their core business based on AI features but do not want to allocate too much budget to build their unstructured data pipeline) and AI-empowered companies (who want to extract business intelligence from their massive unstructured data but don’t want to allocate too much budget to build their unstructured data pipeline) — no matter if they are AI-capable or non-AI capable,” he elaborates.
Currently, Instill AI serves customers in the fields of drones, service robots, cloud security cameras, manufacturing, AI content production, and so on. The company will release VDP under the open-source Apache licence 2.0 to benefit a broader community.
In the meantime, Instill AI offers Instill Cloud, a fully managed cloud service of VDP. The product will be launched early this year. The goal is to serve the community members who want to explore, process or analyse their unstructured data without worrying about the infrastructure maintenance themselves.
Funding and plans
Instill AI recently raised a US$3.6 million seed round of investment from investors such as RTP Global, Lunar Ventures, and Hive Ventures. It will use the money to double the team size by the end of 2023 to build the open-source infrastructure for unstructured data.
It will also continue to improve the user experience for each AI and Data practitioner who works on unstructured data or builds AI-first applications. “We will release a new user dashboard for monitoring, logging and auditing, a new component for logic operators to flexibly manipulate the dataflow, a new drag-and-drop UI to assemble components into pipelines easily, and more data connectors for unstructured data,” Chang explains.
Also Read: Preparing for the AI revolution: Ensuring a positive outcome for humans
The company is also keen to make the model import and deployment more user-friendly. Many new AI tasks will be added, including tasks for Generative AI. To unleash the full power of VDP, model training and evaluation features are planned in the 2023 roadmap.
The market size and trends
According to Precedence Research, the estimated global AI market size was US$87.04 billion in 2021, and it is expected to hit US$1.6 trillion by 2030, with a registered CAGR of 38.1 per cent from 2022 to 2030.
A MarketsandMarkets report pegs the global big data market revenue to be worth US$162.6 billion in 2021, which is poised to reach US$273.4 billion by 2026, growing at a CAGR of 11 per cent from 2021 to 2026.
The modern data stack for unstructured data in the AI and Data market is valued at around US$87.04 billion and $162.6 billion, respectively, in 2021.
Chang observes primarily two trends in the AI industry.
1) No-/low-code AI solution: no-code for the non-tech savvy people to benefit from AI without coding and low-code for the developers to integrate with the existing stack easily.
2) Open platform: MLOps tools have been prosperously developed, particularly for the current best practice of Software 2.0 and data-centric AI. Considering the complexity of the MLOps cycle, it is challenging to build the components of an AI system all from scratch. Instead, people would prefer AI tools/services with vendor-agnostic and open frameworks that ensure easy integrations into the existing tech stack.
The advent of Generative AI such as ChatGPT, DALL-E and Midjourney has shown the power to automate content creation. It is just the beginning. More and more AI tools will emerge to take advantage of the rapid development of technology to find and solve new use cases.
“The AI industry is experiencing a shift from vertical AI applications/products to horizontal AI infrastructure. Machine learning and AI should be as easy to access as other off-the-shelf cloud services in the software industry today. With AI being the infrastructure that transforms every aspect of our lives, ‘AI-first’ will become the default norm,” concludes Chang.
—
Fundraising or preparing your startup for fundraising? Build your investor network, search from 400+ SEA investors on e27, and get connected or get insights regarding fundraising. Try e27 Pro for free today.
The post Instill AI can convert your unstructured data into meaningful data using low-code tools appeared first on e27.