AI Security
Bootcamp

Training professionals to shape how we secure frontier AI systems.

Apply Now
Participant stories from AISB in London. English captions are included in the film.

Past participants have been affiliated with

About

As AI systems become more capable and integrated into critical infrastructure, new attack surfaces and failure modes are emerging that traditional security training doesn't cover. We bring together experienced security professionals and equip them with the threat models, techniques, and hands-on skills needed to engage with the most pressing AI security challenges.

Each cohort is small and intensive, designed so peer learning between practitioners is a meaningful part of the experience. Participants come from offensive security, incident response, threat intelligence, infrastructure, and application security backgrounds; the AI-specific material we cover pushes that experience into new territory.

We run cohorts in different cities, partnering with local AI safety and security organisations. Our programs are free to attend and fully funded for accepted participants.

Cohorts

Upcoming

AISB San Francisco 2027

7-day intensive · 20 participants · January 2027

An AI security cohort in the Bay Area, home to the frontier AI labs. Threat modelling, adversarial attacks, LLM and infrastructure security.

Read more →

AISB London 2026

7-day intensive · 20 participants · December 2026

An intensive London cohort for security professionals shaping how we secure emerging AI systems, from adversarial attacks and LLM security to infrastructure and governance.

Read more →

AISB San Francisco 2026

7-day intensive · 20 participants · October 2026

An AI security cohort in the Bay Area, home to the frontier AI labs. Threat modelling, adversarial attacks, LLM and infrastructure security.

Read more →

AISB Vegas 2026

7-day intensive · 20 participants · August 2026

A frontier AI security cohort during the Las Vegas summer security calendar. Threat modelling, adversarial attacks, LLM and infrastructure security.

Read more →

AISB London 2026

7-day intensive · 20 participants · August-September 2026

An intensive cohort for security professionals shaping how we secure emerging AI systems, from adversarial attacks and LLM security to infrastructure and governance.

Read more →

AISB Singapore 2026

7-day intensive · 16 participants · April 2026

A focused practitioner cohort run alongside Black Hat Asia. Threat modelling, adversarial attacks, LLM and infrastructure security.

Read more →

AISB London 2025

4-week intensive · 20 participants · August 2025

The first AISB cohort. Four weeks of security fundamentals, infrastructure, and AI-specific threats — culminating in a week of capstone projects.

Read more →

Past

Past cohorts will appear here once they conclude.

Interested in a future cohort? Submit an expression of interest and we'll keep you in the loop.

Team

AISB is led by a small core team that designs the curriculum and runs the program across editions. See the full staff list for everyone who has helped run AISB cohorts.

Pranav Gade

Program Director

Pranav Gade

Created AISB to bridge AI safety and security; previously research engineer at Conjecture. Leads curriculum design and program direction.

Jan Michelfeit

Security Lead

Jan Michelfeit

Research engineer at the UK AI Security Institute. Advises AISB and designs the content, drawing on 10+ years securing complex systems and ML infrastructure.

Rhita Ameziane

Operations Head

Rhita Ameziane

AI red teamer, now focused on managing operations across AISB cohorts. Previously a data scientist and engineer at leading consumer-tech companies.

Priyanshu Upadhyay

Founder's Office

Priyanshu Upadhyay

Technical generalist building alongside Pranav & Rhita to make AISB cohorts run smoother and hit harder. Former fintech PM.

See full staff →

FAQs

Ready to Apply?

Applications are open for AISB San Francisco, Jan 10-16, 2027. The deadline to apply is November 7, 2026 (AoE).

Reach out to hello@aisb.dev for any questions about the program.