עדיין מחפשים עבודה במנועי חיפוש? הגיע הזמן להשתדרג!
במקום לעבור לבד על אלפי מודעות, Jobify מנתחת את קורות החיים שלך ומציגה לך רק משרות שבאמת מתאימות לך.
מעל 85,000 משרות • אלפי משרות חדשות בכל יום
חינם. בלי פרסומות. בלי אותיות קטנות.
DOCX, DOC, PDF - בעברית או באנגלית
כבר העליתם קו"ח בעבר? כניסה
The role
Frontier AI agents can now identify vulnerabilities, chain exploits, and operate on live systems. Understanding how they attack and designing realistic tests is critical.
At Innodata’s Agentic Evaluation and Safety practice, you will work with leading frontier AI labs to evaluate offensive cyber capabilities and misuse risks. As Principal Offensive Security Engineer, you will set the technical standard for testing AI agents, design attack scenarios, and partner with safety and alignment researchers to translate complex research questions into rigorous evaluations.
Reporting to the Head of AI Cyber Agentic Environments & Testing and the CTO’s Office, you will help build a team at the forefront of AI cyber security.
What you'll do:
- Design how agents are attacked and tested. Define the attack paths, misconfigurations, and exploit chains that AI agents encounter. Model them on how real enterprises are actually breached.
- Build vulnerable-by-design systems. Author realistic targets across web apps, identity, Active Directory, cloud, and CI/CD, with the depth and texture of real production environments.
- Test frontier agents yourself. Run agents against what you build, study how they reason and act, and find where tests are too easy, too artificial, or leak that they are tests.
- Make results trustworthy. Define what counts as a successful exploit, an unauthorized action, or a data leak, and work with engineering on the signals that capture it.
- Collaborate with alignment scientists. Translate questions about misuse, unsafe autonomy, and agentic risk into concrete offensive scenarios.
- Set the bar for the team. Review scenarios from other engineers, mentor offensive engineers and range builders, and help hire them.
- Represent the work to labs. Walk senior technical clients through findings and methodology.
What you bring
- 8+ years of hands-on offensive security: penetration testing, red teaming, exploit development, or vulnerability research.
- Expert-level web and application exploitation, plus strong depth in at least two of Active Directory and lateral movement, cloud attack paths, or container and Kubernetes security.
- A track record of building things that are meant to be broken: CTF challenges, training labs, ranges, or adversary emulation environments.
- Comfort writing code and infrastructure: Python or Go, Docker, Linux, networking.
- Hands-on experience with LLMs or AI agents in a security context, such as red-teaming models, building AI-driven offensive tooling, or evaluating agents on security tasks.
- Clear thinking and writing. You can explain an attack chain to a researcher, an engineer, or a client.
- You raise the level of the people around you.
- Mentor junior SWEs, Cyber Researchers, and RL Engineers
Nice to have
- Background in elite offensive units, national CERTs, or top-tier red-team consultancies.
- Published CVEs, Hack The Box or similar machines, or strong CTF results.
- OSWE, OSEP, OSEE, CRTO, or equivalent demonstrated skill.
- Familiarity with AI safety and alignment research: misalignment, evaluation awareness, agentic risk.
- Talks or publications at Black Hat, DEF CON, or academic security and AI venues.
We’re an equal opportunity employer.
Only relevant applicants will be contacted
במקום לעבור לבד על אלפי מודעות, Jobify מנתחת את קורות החיים שלך ומציגה לך רק משרות שבאמת מתאימות לך.
מעל 85,000 משרות • אלפי משרות חדשות בכל יום
חינם. בלי פרסומות. בלי אותיות קטנות.
DOCX, DOC, PDF - בעברית או באנגלית
כבר העליתם קו"ח בעבר? כניסה