Việt Anh Nguyễn

Reviewed AI-generated Vietnamese audio and text content for quality and safety compliance, identifying harmful or policy-violating segments and providing structured feedback to support filtering and re-training decisions. Audited labeler outputs against structured rubrics, logged safety findings, and supported calibration to reduce inter-annotator variance in ambiguous categories.

Reviewed Vietnamese-language AI-generated content against detailed annotation guidelines for accuracy, cultural appropriateness, and adherence to project-specific safety and quality standards. Applied precise policy interpretation to ambiguous cases, documented flagging rationale, and adapted to evolving guidelines and safety taxonomies in an asynchronous multilingual team environment.

Interpreted and applied complex legal policies and regulatory frameworks with evidence-based precision, including structured research into policy gaps and compliance edge cases. Reviewed and validated large volumes of bilingual (Vietnamese/English) documentation while maintaining confidentiality and data security standards for sensitive materials.

Conducted systematic LLM red teaming by designing adversarial prompts and documenting evasion techniques, normalization patterns, and failure modes across hate/harassment, self-harm, sexual content, misinformation, and malicious instruction domains. Annotated and scored model responses against multi-dimensional safety rubrics in English and Vietnamese, including intent ambiguity and escalation ris

Grade: GPA 3.32/4.0 (Distinction)

Activities and societies: Second Prize - Best Thesis Award (2024).

Earned a Bachelor of Laws (LL.B.) with a GPA of 3.32/4.0 (Distinction). Completed coursework in constitutional, criminal, international, and comparative law, and received a Second Prize for the best thesis (2024).