AIM Intelligence Stinger Red Teaming and Starfort Guardrails

AIM Intelligence Co., Ltd. is a Seoul-based AI security startup founded in 2024 and led by CEO Sangyoon Yu. It works to keep generative AI secure, aligned and controllable, and its research team has published at ICLR, ICML, ACL and NeurIPS. It has served as an APAC AI safety partner for OpenAI and works with Korean banks, telecoms and electronics groups.

Stinger is an automated red teaming product that generates attack scenarios across text, image, audio, video and physical AI, including agentic red teaming and custom tests based on business logic. Starfort provides low-latency, policy-aware guardrails that block malicious prompts, detect and mask personal data and trade secrets, and control abnormal agent API calls. Both work with proprietary LLMs, custom agents, commercial APIs such as ChatGPT and Gemini, and coding agents, and can run as cloud SaaS or fully isolated on-premise.

• Stinger red teaming: auto-generated attack scenarios for text, image, audio, video and physical AI.
• Agentic red teaming: tests tool-using agents and business-logic abuse.
• Starfort guardrails: low-latency blocking of malicious prompts and unsafe outputs.
• Data protection: detects and masks personal data and trade secrets.
• Agent control: flags and blocks abnormal agent API calls.
• Benchmarks and consulting: safety benchmarks for LLM, VLM and VLA models plus security advisory.

Vietnamese banks and card issuers can red-team customer chatbots and agents before launch and add guardrails that mask account data. Telecom operators can protect public AI services from jailbreaks. Korean-invested manufacturers in Vietnam can test robotics and physical AI for unsafe behaviour.

Security teams find jailbreaks and unsafe agent behaviour automatically, across text, image and audio, before attackers do. Guardrails mask personal data and trade secrets in real time, reducing leakage risk when staff and customers use GenAI. One vendor covers both offensive testing and runtime defence, closing the loop between findings and fixes. On-premise deployment suits banks and telecoms with strict data rules.

Banks and telecoms in Korea and across Asia are putting LLM assistants and agents in front of customers, and regulators increasingly expect red teaming and guardrails. Vietnamese financial institutions and telecom operators are following the same path and need multilingual, on-premise AI security tools.

Founded in 2024 in Seoul. Investors listed include Samsung Venture Investment, Mirae Asset Capital, Smilegate Investment, LG U+ and Z Venture Capital, with a Pre-A of 1.3 million USD in August 2025, a Series A reported in April 2026 and investment from LG Electronics and Naver D2SF in June 2026. Won the Supernova Challenge at GITEX Global 2025. ISO 27001 certified.