Offensive Security Engineer

ClickHouse ClickHouse · Data AI · EMEA · Security

Offensive Security Engineer role focused on identifying and mitigating security vulnerabilities in ClickHouse's offerings, with a significant emphasis on assessing and attacking AI/LLM-specific features and tooling. The role involves building and operating agentic and LLM-assisted tools for offensive security use cases, including reconnaissance, exploit chaining, and fuzzing, and partnering with detection engineering to validate controls.

What you'd actually do

  1. Assess AI/LLM-specific attack surface across ClickHouse's own AI-powered features and internal AI tooling, including prompt injection, model/data exfiltration, and unsafe agentic tool-use patterns
  2. Build and operate agentic tooling for reconnaissance, exploit-chaining and attack-path discovery, and apply LLM-assisted fuzzing to accelerate vulnerability discovery across ClickHouse's products and infrastructure
  3. Identify security gaps and vulnerabilities in all ClickHouse offerings triage a wide range of vulnerabilities reported via our bug bounty program, responsible disclosure, GitHub Issues covering web, API and server - client assets including low level memory issues like heap or buffer overflows
  4. Improve and develop security assurance activities - pentests, vulnerability assessments, bug bounty programs, fuzzing
  5. Plan and execute internal red team assessments and penetration tests against ClickHouse infrastructure, cloud environments, designing realistic adversary scenarios to test detection, response and control effectiveness

Skills

Required

  • 7+ years of experience in pentesting, red teaming and product security
  • Experience supporting engineering and product implementation efforts by performing threat assessments, assurance activities, advisory as well as, in some cases, implementation work across distributed systems covering web, API, client/server assets
  • Hands-on experience conducting offensive security engagements - internal red teaming, penetration testing and adversary simulation - across cloud, network and application environments
  • Ability to design adversary scenarios grounded in real threat intelligence (e.g. ransomware operators, supply chain attackers, insider threat) tailored to ClickHouse's risk profile as a multi-tenant cloud data platform
  • Strong written and verbal communication skills, with ability to translate attack chains into clear, actionable findings for engineering and leadership
  • Experience building or adapting agentic and LLM-assisted tooling for offensive security use cases (recon automation, exploit chaining, fuzzing harnesses), with judgment on where AI genuinely speeds up an engagement versus adds noise
  • Familiarity with AI/LLM-specific vulnerability classes and testing methodology
  • Strong knowledge of and experience with one or more cloud service providers (e.g. AWS, GCP, Azure), Kubernetes, Cilium
  • Experience implementing and operating engineering security tools and processes (e.g. static / dynamic code analysis, software composition analysis, SBOM, OWASP SAMM, client and network fuzzing tools)
  • Security as code mindset, with focus on solving problems with automation and scale in mind

Nice to have

  • BS, MS, or PhD in Computer Science or related field
  • Previous contributions to open source projects
  • Security or cloud related certifications (AWS, GCP, Azure)
  • Previous experience in using AI security harnesses for pentesting purposes
  • Experience building or operating internal red team tooling and infrastructure from scratch, rather than relying solely on off-the-shelf frameworks
  • Offensive security certifications (e.g. OSCP, OSCE, OSEP, OSWE)

What the JD emphasized

  • AI/LLM-specific attack surface
  • agentic tooling
  • LLM-assisted tooling
  • agentic and LLM-assisted tooling for offensive security use cases

Other signals

  • AI/LLM-specific attack surface
  • agentic tooling for reconnaissance
  • LLM-assisted fuzzing