Cybersecurity and AI engineering
We test and fix the security of your software. Use AI safely, without exposing your company's data.
Web application pentests, cloud reviews, and security for AI applications, run by a senior team. Plain-language report, scope and price fixed in writing before we start, and remediation by our engineers when it is part of the project.
What we do
Offensive security and AI engineering.
Security
Security Assessment
External attack surface, a pentest of one web application or API, and a review of one cloud account. Report with one retest included.
2 weeks
Fix and Harden
We fix findings through pull requests, add security checks to your CI, and clean up access, MFA, and backups.
Monthly, no lock-in
Security for AI Applications
Testing for prompt injection, data leakage, and tool permissions in chatbots, agents, and RAG systems.
2 weeks
AI engineering
AI automation and applications
Assistants, document processing, and internal tools, integrated with the systems you already use. Every project starts with an assessment of where AI pays off and where it does not.
Models and data
Forecasting and classification models trained on your data and tested against your numbers. Dashboards for the questions your team asks every week.
How we work
- 01
Scope and proposal
A scoping call with the technical team. A proposal with scope, timeline and price in writing.
- 02
Execution
Manual testing led by senior specialists, following the OWASP WSTG and PTES. One point of contact follows the project from start to finish.
- 03
Report and walkthrough
A plain-language report, reviewed before delivery, and a meeting to walk your team through every finding.
- 04
Fix and retest
Your team fixes, or ours does through pull requests. The retest confirms the result.
See what your company exposes to the internet
Give us your domain. Within 2 business days you get a summary, reviewed by our team, of what we found: subdomains, open ports, certificates, email without spoofing protection, and leaked credentials. No cost and no intrusive testing.