Guarantee Secured
Automotive
AI Systems

Eliminate the threat of injected prompt triggering information about violence, profanity, criminal behaviours, race, or other unsavoury topics. Powered by the world’s most advanced AI defence technology.

OUR PRODUCTS CONTACT US
Our Trusted Partner

Here at ZEROAI we focus on markets where technology, innovation, and capital can unlock long-term value.

Become our partner>
Adidas
Chanel
Nike
Toyota
GS1
BlackBerry
Mini
Honda
Sentinel & DriveGuard

Our advanced defence tech for input & output of large scale model.

Input Level Defence

Based on Stackelberg Reinforcement Learning Framework, our sentinel model perform filtering of inputs to achieve protection against the toxic input.

Output Level Defence

DriveGuard is able to prompt human intervention. With redundant autopilot system, it intervene while deeming the output of current model to be unsafe.

Safe Autonomous Driving

Lightning fast! Our real time-inference, safety assured automotive large model.

10×

Faster LLM Aided Inference
We introduce an innovative framework that trains an offline RL based on synthesized data through LLMs. This methodology centres around utilizing a dataset generated by an LLM-based agent operating offline. Boosting the inference speed of the agent by 10× with knowledge from LLMs.

  1. LLM Assisted System

    The large model adopts the ReAct strategy, which involves a cycle of self-reflection, action, and observation, to perceive and analyse its surrounding environment.

  2. Visual Language Model

    Integrating digital vector modality into the pre-trained LLMs. This type of data is widely used in robotics technology, representing information such as speed, actuator positions, and distance measurements.

  3. Faster Decoding Procedure

    We incoperate assisted decoding statgey, inspired by speculative sampling, to accelerate the entire system with mathematical gurantee on generation quality.
AI Evaluation

Ensuring consistency and security of LLM through the full lifecycle.

Only by ensuring consistency and security can large language models be cautiously deployed in specific application scenarios, thereby effectively promoting the automation and intelligent decission-making processes of tasks.

Knowledge and Ability Assessment
The goal is to provide a comprehensive and systematic overview covering recent advances in basic knowledge and reasoning capabilities of large language models and their corresponding benchmarks
Consistency Supervision
Focuses on examining the performance of LLM along key dimensions, including ethics and moral assessment, bias detection, toxicity assessment, and authenticity assessment
Safety Supervision
The evaluation and regulation of LLM will be completed independently by AI model, achieving fully automated generation of safety assessment supervision reports
Advantages

Top background, top experience, top products, internationalized

Top academic background team

Our technical team, comprised entirely of Ph.D. holders from top 100 global universities, collaborates with esteemed AI security professors and large model postdoctoral fellows. Specializing in areas like autonomous driving and deep learning, we leverage cutting-edge research for technological innovation.

Rich industrial experience

Our team has substantial industry experience, with members working in leading firms like Alibaba, Huawei, DJI, and TikTok, and others being serial entrepreneurs. This diverse experience allows us to effectively turn research into practical applications and solutions.

Entrepreneurship and financing experience

We have solid entrepreneurial and financing skills. Some members have successfully started and profited from businesses, while others have worked in venture capital. This experience gives us a keen insight into market needs and business development.

Multiculturalism and diverse internationalization

The team is multicultural and internationally diverse, with educational and life experiences from the UK, China, the US, Germany, Canada, and more, with over 40% female representation. This diversity offers us a wide perspective in innovation and problem-solving.

Our People Make Us Great

Motivated researchers, engineers, IT ops and business teams collaborating at high velocity.

William Wu

CEO

Gerald Yin

Research Leader

Tudor Jianu

Researcher

Josh

Director

Xiaoyi Zou

COO

Jianhong Wang

Scientific Advisor

William Wu

CEO

Gerald Yin

Research Leader

Tudor Jianu

Researcher

Josh

Director

Xiaoyi Zou

COO

Jianhong Wang

Scientific Advisor

CONTACT

GET IN TOUCH

Want to know more about product feature? Need details about our Business plan? Interested in join us? Let us know.

 Digital Innovation Facility, Dover St, Liverpool L69 3RF.
 +44 0000000000
 zeroai2024@gmail.com

Please fill out this field.

Please fill out this field.