Cloudsine is a cloud technology company specialising in cloud adoption, security and innovation. Cloud will be one of the most significant technology shifts organisations will encounter in the next few years. Hence Cloudsine aims to be the trusted partner to help customers build new capabilities and business opportunities for the digital future. Since 2012, cloud, security and innovation have been the DNA of their team.

They have built cloud systems to support numerous large enterprise and government customers across APAC countries, including Singapore, Australia, Philippines, India, Indonesia, Hong Kong, etc. Their cloud experience includes web applications, content delivery, WAF, encryption, logs, analytics, containers, serverless architecture, deep learning, storage and archival, etc. In addition, they have their own Intellectual Property and tools (eg. WebOrion security suite) to complement existing public cloud tools to help customers strengthen their cloud security.

Project role: AI Security Engineer (Project Rhythm)

This is an experimental project where Cloudsine explores various latest AI technologies such as ChatGPT, Large Language Models (LLMs), etc. and explore various attack vectors and ways to secure them. Project Rhythm aims to provide assurance and privacy to organisations concerned about the privacy, confidentiality and integrity of AI services. Trainees will use Jupyter notebooks, Pytorch, Tensorflow and run some of the AI models in AWS cloud or via API calls.

The project could also expand into looking at implementing LLMs on edge devices and test out the feasibility, performance and security of these devices. Trainees are expected to be very hands-on and be responsible for researching the latest AI models, e.g. the LLM, Llama2, ChatGPT, and their potential security vulnerabilities. They will also develop demos and prototypes to showcase security vulnerabilities, e.g. prompt injection attacks, and test and implement POC solutions to address the above security vulnerabilities, which can be built on Bootstrap, Python, Flask, etc., on AWS. Proficiency in Python, Pytorch or Tensorflow will be required.