AI Model Security, Trust, and Privacy Protection

STYRK focuses on AI Security and risk management, offering tools to measure, monitor, and mitigate vulnerabilities, bias, and adversarial threats across machine learning and large language model systems. Its AI Portal provides continuous scanning for adversarial behavior, runtime input validation, and contextual guardrails that enforce secure and trustworthy model use.

The company’s capabilities span adversarial training, model hardening, anomaly detection, and comprehensive risk assessments aligned with frameworks such as NIST AI RMF, ISO/IEC 42001, and the EU AI Act. STYRK places particular emphasis on safeguarding visual AI systems used in mission-critical detection environments, helping organizations maintain performance and reliability under real-world threat conditions.

The company has raised $3.5 million, including a $2.5M seed round (Dec 2024) led by SmartFin Ventures and a $1M pre-seed round (May 2024) backed by Punja Global Ventures.