Available for engagements - Brașov, Romania
Penetration Tester @ NTT DATA
Offensive Security · Red Team · Cloud & AI Security
I'm Petre Radu Cătălin, a Penetration Tester at NTT DATA. I identify critical vulnerabilities in Active Directory, cloud environments, and complex web applications - then deliver remediation guidance that security teams can act on immediately.
// whoami
Offensive security professional · Penetration Tester at NTT DATA · Based in Brașov, Romania. Known as Petre Radu to clients, peers, and the security community.
I'm Petre Radu Cătălin - often simply known as Petre Radu - an offensive security professional and Penetration Tester at NTT DATA, based in Brașov, Romania. Every day I probe enterprise networks, web applications, and cloud infrastructure for the flaws that attackers are paid to find, and I translate them into risk owners can understand and engineers can fix.
My approach to security is simple: real risk over noise. Anyone can produce a spreadsheet of low-severity findings. My work focuses on what actually matters - the vulnerabilities that persist, the paths that lead to domain compromise, and the remediation guidance that a security team can act on the same day. Every report I ship includes a clear, prioritized, and technically accurate plan to get from 'vulnerable' to 'defended'.
My path runs through the Master's programme in Cyber Security at Universitatea Transilvania din Brașov (2024-2026), where I studied advanced cyber defense, red and blue team operations, and security auditing - and where I complemented the curriculum with Google.org Cybersecurity Seminars in 2025. That foundation carried me into enterprise penetration testing at NTT DATA, where I lead assessments against production systems at scale.
Between full-time engagements, I work as a Bug Bounty Hunter on Intigriti and as a Vulnerability Researcher on HackerOne, hunting for flaws in production systems through responsible disclosure. This freelance work keeps me sharp: real bug bounty programs are a constant grind of edge cases, business logic abuse, and access control failures that no lab environment can reproduce.
My scope extends past traditional pentesting into AI security evaluations - testing LLM applications for prompt injection, data exfiltration, and excessive agency - and into the regulatory world through GDPR assessments and NIS2 implementation support. Security is a moving target, so I treat continuous learning as part of the job: new platforms, new tools, new attack classes, every single week.
// capabilities
Four interlocking disciplines. My offensive work is grounded in an understanding of defense - every assessment I run, from web penetration testing to cloud security review, is built to produce actionable defense.
Hands-on exploitation across the full attack surface - from web apps to the domain.
Breaking and hardening modern, ephemeral attack surfaces before attackers do.
Turning regulatory requirements into measurable, defensible security posture.
Knowing the defender's playbook sharpens every offensive engagement.
// history
Enterprise penetration testing, bug bounty hunting, and a Master's in Cyber Security - a career built on finding real vulnerabilities and fixing them for good.
Penetration Tester
Nov 2025 - PresentRomania · Remote
Bug Bounty Hunter
May 2026 - PresentRemote
Vulnerability Researcher
May 2025 - Sep 2025Remote
Master's, Cyber Security
2024 - 2026
// featured_work
Selected projects from research, development, and real engagements - static malware analysis, forensic tooling, and the automation that powers my assessments.
PE File Analyzer for Static Malware Analysis
MalwarePeek is a Python-based portable executable analyzer built for malware analysts and reverse engineers. It parses PE headers, fingerprints packers, and scans samples against YARA rules - then generates clean HTML reports for triage.
Educational PoC - Authenticode Signature Extraction
A focused educational proof-of-concept that inspects authenticated code-signing metadata inside signed PE files. It demonstrates how signature blocks remain queryable and how tooling can surface embedded certificate chains for forensic review.
PrivateRecon & Vulnerability-Scanning Automation
An internal toolkit of Python and Bash scripts that automate reconnaissance, endpoint enumeration and vulnerability scanning. Used to deliver faster, repeatable assessments and cut end-to-end testing time by roughly 30% at NTT DATA.
// credentials
100+ certifications earned across security, cloud, and AI disciplines - plus CTF results that put the theory to work.
Curated highlights below - full list available on request.
TryHackMe Junior Penetration Tester (PT1)
TryHackMe
OpenAI Cyber Practitioner
OpenAI · PartnerU
Microsoft AI Skills Fest 2026
Microsoft
AZ-700: Azure Networking Solutions
Microsoft Certified
Teaching the AI Fluency Framework
Anthropic
Puppet Environment Lab
Skillsoft
Google.org Cybersecurity Seminars
Google.org
Offensive security is a competitive sport. These results are how I train between engagements.
HackTheBox Season 9
Holo Tier
TryHackMe Industrial Intrusion CTF
Team Rank 6 / 20,220
TryHackMe Honeynet Collapse CTF
Solo Rank 66 / 960
// arsenal
The tools I reach for daily - offensive platforms, cloud CLIs, and the defensive stack I know well enough to test around.
Cobalt Strike
Sliver
Nighthawk
BloodHound CE
SharpHound
Rubeus
KrbRelayUp
NetExec
Impacket
Responder
Evil-WinRM
Ligolo-ng
Burp Suite Pro
Nuclei
ffuf
Pacu
AzureHound
GCPwn
Prowler
Garak
// write_ups
Technical write-ups on offensive security, cloud threats, and AI security - the lessons from real engagements, published to help the community.
A field guide to how I evaluate AI systems offensively - prompt injection, indirect injection through retrieved content, and the data-exfiltration bugs that leak what LLMs are told through their own outputs.
The top cloud misconfigurations I encounter on real engagements - S3 buckets, over-permissive roles, exposed credentials, and public snapshots - and how to find them before the attackers do.
A practical, phase-by-phase breakdown of how a single web application flaw can be chained into full Active Directory compromise - and the detection gaps that let it happen.
// connect
Available for remote and on-site engagements. Based in Brașov, Romania.