Aizhan Azhybaeva

Aizhan Azhybaeva is a GenAI application QA engineer at genai.qa - the specialist practice for red-teaming and safety-testing GenAI applications.

Focus

Aizhan publishes research and practitioner guidance on:

• GenAI application red-teaming - systematic adversarial testing of LLM-powered products and AI agents
• Prompt injection testing - direct and indirect injection attack patterns, OWASP LLM Top 10 categories
• RAG evaluation - retrieval quality, context faithfulness, grounded answer validation
• Agent safety testing - tool-use safety, autonomous agent behaviour, guardrail validation, OWASP Agentic AI Top 10
• Jailbreak surface mapping - refusal quality, alignment assessment, multi-turn manipulation
• Hallucination benchmarking at the application layer - distinct from model-layer measurement
• Multi-modal red-teaming - vision, audio, and cross-modal adversarial testing
• Agentic system exploitation - tool poisoning, memory manipulation, privilege escalation through tool chains

genai.qa is the application-layer counterpart to aiml.qa (model layer). aiml.qa tests ML models and evaluation; genai.qa tests GenAI applications and red-teams them. In the NomadX AI lifecycle: Design (mlai.qa) -> Validate (aiml.qa) -> Stress-Test (genai.qa) -> Grow (generative.qa).

Topics Covered

• GenAI Application QA
• Application-Layer AI Red-Teaming
• Prompt Injection Testing
• Jailbreak Surface Mapping
• RAG Evaluation
• Agent Safety Testing
• Agentic AI Top 10
• OWASP LLM Top 10
• Hallucination Benchmarking
• AI Safety Alignment
• Multi-Modal Red-Teaming
• Tool Poisoning Detection
• Memory Manipulation Attacks
• GenAI Compliance QA

Profiles & Links

• https://www.linkedin.com/company/nomadx-ae
• https://github.com/nomadx-ae