NinjaTech AI, a San Francisco-based startup providing conversational artificial intelligence for business, announced the public beta launch of its new AI agent service, Ninja. Powered by purpose-built machine learning (ML) chips from Amazon Web Services (AWS), Ninja goes beyond traditional AI assistants and co-pilots by bringing autonomous agents to market.
Ninja has the ability to asynchronously plan and execute real-world tasks, such as scheduling meetings on your behalf, assisting with coding tasks, and conducting in-depth research. The AWS chips Trainium and Inferentia2, along with Amazon SageMaker, a fully managed cloud-based ML service, enable Ninja to create, train, and scale custom AI agents that can autonomously manage complex tasks.
Leveraging AWS cloud capabilities, Ninja users can assign new tasks without waiting for ongoing tasks to be completed. The service notifies users when an operation is completed or if there are questions that require input. Users can also click on the sidebar to see the progress of each task.
“Working with AWS Annapurna Labs has been a true game changer for NinjaTech AI. The power and flexibility of the Trainium and Inferentia2 chips for our reinforcement learning AI agents far exceeded our expectations. It is easy to integrate and elastically scales to thousands of nodes via Amazon SageMaker,” said Babak Pahlavan, founder and CEO of NinjaTech AI.
Pahlavan further stated that these new AWS-designed chips can reduce total costs by up to 80% and are 60% more energy efficient compared to similar GPUs on the market. The chips also offer native support for the larger 70B variants of the latest open source models such as Meta's Llama 3.
AI agents rely on LLMs that are tuned for specific tasks through extensive training and fine-tuning, such as reinforcement learning. However, the scarcity and inelasticity of GPUs required to train LLMs poses a key challenge to the successful deployment of AI models. AWS provides a solution by offering a combination of custom chip technology and scalable cloud infrastructure.
(Ole.CNX/Shutterstock)
The four main conversational AI agents offered by Ninja include Real-Time Web Search, Ninja Coder, Basic Scheduler, and Ninja Advisor. Additionally, Ninja offers limited access to third-party LLMs, allowing users to compare side-by-side results from leading LLMs from companies like Google, Anthropic, and OpenAI. The free version of Ninja offers only a limited number of daily tasks. To access more tasks, users must subscribe to a paid plan.
NinjaTech AI is approaching its goal of providing a service that allows busy professionals to get the most out of AI. Our cutting edge asynchronous infrastructure allows Ninja users to manage a nearly limitless number of tasks simultaneously – it's like harnessing the power of multiple AI models without having to manually manipulate each model individually.
The release of Ninja marks the next generation of productivity tools that can transform the way we work, learn, and collaborate. It also highlights AWS' growing influence in the AI and ML market. NinjaTech AI is one of many companies that have chosen AWS Inferentia and AWS Trainium chips over NVIDIA GPUs.
Related Items
AI chatbots: a hedge against inflation?
OpenAI rival Inflection AI raises $1.3B to power Pi chatbot
University of Michigan is developing an AI coaching bot for students
Related
