Benjamin W.

Senior Software Engineer

Wrocław, Poland

Experience

Jan 2023 - Present
2 years 10 months
Belgrade, Serbia

Senior Software Engineer

Flow Ninja

  • Evaluated LLM and AI agent vulnerabilities using Python and Bash scripts, identifying prompt injection risks and reducing exploitable scenarios by 35%.
  • Developed containerized test environments and CI/CD security pipelines with Docker, accelerating reproducible evaluation cycles by 28% across multiple AI systems.
  • Implemented automated test harnesses and offline reproducible scenarios, improving assessment coverage for model behavior under adversarial conditions.
  • Led security research initiatives targeting LLM safety, designing attack simulations that enhanced model robustness against malicious input patterns.
  • Collaborated with cross-functional teams to advise on secure deployment practices, integrating AI models into cloud environments without compromising safety.
  • Analyzed network and application-level security, implementing mitigation strategies to harden LLM and AI agent infrastructure.
  • Applied reverse engineering tools including Ghidra to assess AI-related software components, uncovering critical security gaps and potential exploits.
  • Contributed to internal documentation of AI red-teaming best practices, improving knowledge transfer and onboarding efficiency.
  • Monitored and optimized cloud and containerized deployments to ensure secure, reliable, and scalable AI operations across distributed environments.
Jan 2019 - Dec 2022
4 years
Cluj-Napoca, Romania

Senior Software Engineer

BetterQA

  • Constructed Python and Bash automation scripts for AI model testing and vulnerability scanning, increasing testing throughput by 30%.
  • Designed reproducible test scenarios for AI agents, enabling offline evaluation of model safety and security compliance.
  • Implemented Docker-based CI/CD pipelines for secure deployment of AI and ML workloads, reducing human error and ensuring auditability.
  • Performed penetration testing across web, API, and infrastructure components, identifying critical vulnerabilities and enforcing mitigation strategies.
  • Collaborated with cross-functional teams to integrate AI safety best practices into development workflows and deployment processes.
  • Developed monitoring solutions for AI model deployments, detecting anomalous behavior and potential security threats.
  • Advised product teams on secure coding practices and vulnerability mitigation, improving overall AI system resilience.
Mar 2017 - Dec 2018
1 year 10 months
London, United Kingdom

Software Engineer

Valor Software

  • Implemented Python and Go backend services with security-focused workflows, enhancing reliability and maintainability for AI infrastructure.
  • Conducted penetration tests and security evaluations on backend services, reducing exploitable vulnerabilities by 25%.
  • Automated repetitive testing tasks with scripts and CI/CD integration, improving development efficiency and reducing manual errors.
  • Analyzed cloud-based deployments and network configurations, improving secure connectivity and system reliability.
  • Collaborated with DevOps teams to integrate security checks into deployment pipelines, enforcing compliance and best practices.
Aug 2015 - Feb 2017
1 year 7 months
Warsaw, Poland

Software Engineer

Smartym Pro

  • Engineered backend automation scripts for security and performance monitoring, increasing operational efficiency across multiple AI services.
  • Performed vulnerability assessments and Linux system audits, reducing security risk and hardening server environments.
  • Developed containerized environments for AI model deployment, improving reproducibility and team collaboration.
  • Implemented CI/CD pipelines with integrated security checks, ensuring reliable and secure software releases.

Summary

Evaluated AI/LLM models for vulnerabilities and safety risks, improving model reliability and reducing exploitation exposure by 35% across test scenarios. Developed automation scripts, test harnesses, and reproducible evaluation pipelines using Python and Bash, accelerating AI red-teaming cycles by 28%. Implemented containerized CI/CD security workflows with Docker, ensuring secure model deployments and reliable integration in distributed environments.

Languages

English
Advanced
Polish
Intermediate
Romanian
Intermediate
Serbian
Elementary

Education

Oct 2012 - Jul 2015

University of Warsaw

Bachelor’s degree · Computer Science and Technology · Warsaw, Poland

Need a freelancer? Find your match in seconds.
Try FRATCH GPT
More actions