Contract Developer Jobs in Kenosha, WI (NOW HIRING)

About Cobalt:

Cobalt is seeking security engineers to adversarially test frontier models and AI agents, and to produce the reasoning and evaluation data used to measure and improve how those systems hold up under pressure.

This opportunity is suited to people with a security background who have turned their attention to AI systems, whether professionally, through structured red teaming and model bounty programs, or through sustained independent research. Traditional offensive security experience without prior AI work is welcome, provided you can code and can learn a new target quickly.

You do not need a formal AI research background. What matters is that you can reliably find failure modes others miss, that you can tell a lucky one-off from a reproducible weakness, and that you can build the harness needed to prove it.

All work is performed against sandboxed environments, purpose-built targets, and model endpoints supplied by us or by the lab. We do not accept work performed against systems you are not authorized to test, and we do not accept material covered by a client agreement or obtained without authorization.

What you'll do:

Depending on the project, you may:

• Probe models and agents for failure modes, including prompt injection, unsafe tool and browser use, data exfiltration paths, and breakdowns in refusal or policy adherence, and document each case so another engineer can reproduce it
• Build test harnesses and automation that run attempts at scale against model APIs and agent frameworks, rather than testing by hand
• Produce written attack traces, capturing how you form and test hypotheses, what you rule out and why, and how you arrive at a working approach, rather than only the successful prompt
• Author adversarial evaluation scenarios and benchmark items with verifiable success criteria, including multi-turn and agentic tasks
• Design rubrics and partial-credit criteria for scoring model behavior, and classify observed failures into a consistent taxonomy

Projects follow their own guidelines, scope rules, and quality standards, and you will work with feedback from reviewers and lab research teams.

Required qualifications:

• Demonstrable security experience, and demonstrable adversarial work against language models or AI agents through professional red teaming, structured bounty programs, published research, or a substantive independent record
• Strong coding ability, typically Python, sufficient to build your own harnesses, automate at scale, and work with model APIs and agent frameworks
• Depth in at least one relevant area, for example prompt injection and jailbreak techniques, agent and tool-use security, multimodal attacks, or cyber misuse evaluation
• Ability to distinguish reproducible weaknesses from one-off outputs, and to explain each step of your reasoning clearly in writing
• Willingness to work strictly within defined scope, to follow the project's disclosure process, and to sign a confidentiality agreement covering findings and project materials

Why join Cobalt AI:

• Advance frontier AI where it counts. Apply your security expertise to data that frontier labs cannot obtain any other way, where your judgment directly shapes how the next generation of models reasons about security.
• Grow professionally. Expand your influence through evaluation projects, advisory roles, and research collaborations, while developing a working understanding of how frontier models are trained and assessed.
• Work with a top-tier network. Collaborate with security engineers and researchers from leading organizations on high-impact, flexible work.
• Set your own schedule. Flexible 10 to 40 hour weeks that fit around your existing engagements and your life.
• Competitive pay. Rates vary by project and are determined by a number of factors, including scope, skillset, and experience.