AI Security Intern, Vulcan
Description
About Us
AIFT is a leading firm specializing in innovative solutions to help businesses protect their digital assets. We leverage cutting-edge technologies, including artificial intelligence, to deliver comprehensive security assessments and remediation strategies. Vulcan is a groundbreaking product developed by AIFT.
If you are passionate about making a unique impact, as well as securing the future for enterprises through delivering cybersecurity for GenAI - one of the most disruptive and fastest-growing technology trends, this is the role for you.
Know more about how we deliver transformation at:
About the Role
You help Vulcan stay at the front of AI and agentic AI security. Working closely with our AI security researchers and consultants, with day-to-day guidance and mentoring, you’ll track new attacks, replicate them, write them up, and help turn them into tools the whole team can use. You’ll also support our consultants on PoCs and benchmarks.
We want a doer and a fast learner who uses AI to get more done. You'll ship real work in your first weeks, not watch from the side.
- Commitment: At least 3 days per week
- Duration: Minimum 3 months, with the opportunity to extend based on performance and mutual fit
- Location: Taiwan or Southeast Asia
What You Will Do
1. Understand the big picture
- Build a solid understanding of AI safety and security: how AI and agentic systems are built, deployed and attacked, and why it matters to the businesses that run them.
- Learn what Vulcan does: our red teaming and assessment services, the frameworks we use (OWASP Top 10 for LLM Applications, MITRE ATLAS, NIST AI RMF), and how our delivery loop turns what we learn into tools.
- Shadow live work where possible, so you see how real client systems fail.
2. Assist on research themes
- Support the research themes and focus areas set by the Research Engineer.
- Track new attack techniques, disclosures and papers in those areas, and summarise what matters each week.
- Replicate promising techniques in our lab environment, with guidance, and record what worked and what didn't.
- Write up attack methods clearly enough that anyone on the team can repeat them, and map them into the attack library and our risk taxonomy.
3. Build
- Build and maintain evaluation datasets and attack payload sets, including multilingual ones.
- Write scripts and use AI tools and agents to automate repetitive research and testing tasks.
- Run benchmarks that compare a PoC or tool against the manual method, and report the results.
- Test PoCs before they are handed to the platform, and report bugs and gaps.
What You Bring
- Currently studying, or recently graduated, in computer science, cybersecurity, AI or a related field.
- Basic Python: you can write and debug small scripts.
- Curiosity about how AI systems work and fail, such as prompt injection and jailbreaks.
- A doer's attitude: you try things, ship something, and improve it rather than waiting for perfect instructions.
- Fast learning: you pick up unfamiliar tools and topics quickly and ask good questions.
- AI-native working habits: you use LLMs, coding assistants and agents to research, write and build faster, and you check their output critically.
- Clear writing: you can explain what you found and how to repeat it.
- Fluent Mandarin and English.
Nice to Have
- CTFs, bug bounties, security clubs or personal security projects.
- Experience with LLM APIs, agent frameworks or MCP.
- Arabic or other languages relevant to Asia or the GCC, for multilingual testing.
- A GitHub profile or write-ups that show what you've built.
What Success Looks Like
- Within your first month, you can explain the AI safety and security landscape and how Vulcan's work fits into it.
- A weekly research summary the team actually reads and uses.
- Techniques replicated and written up so others can run them without you.
- Attack library and taxonomy entries that pass review with minimal rework.
- Evaluation datasets and benchmark results used in real PoC decisions.
- At least one repetitive task you automated with AI during your internship.
What You'll Get
- Hands-on work in AI red teaming and agentic AI security, a fast-growing field.
- Mentoring from experienced security researchers and consultants.
- Real ownership: your work goes into the tools and methods our team uses on client projects.
- For those close to graduation, strong performers may be considered for a full-time role after graduating.
Application Process
- If you're ready to embark on this exciting journey and contribute to shaping the future of AI security, please submit your resume outlining your relevant experience and motivation for applying.