Aysha Nasim

Architected and deployed multi-agent AI systems in Azure AI Studio and Azure Foundry to automate legal guidance
interpretation across Legal Aid and Office of the Public Guardian (OPG) platforms.
• Built end-to-end RAG pipelines enabling caseworkers to query complex legal documents in natural language,
significantly reducing manual document review time.
• Developed a GPT-4-powered financial analysis agent with anomaly detection to flag irregularities in OPG financial
records, improving compliance review efficiency.
• Designed LLMOps monitoring dashboards tracking model performance, token usage, latency, and system health
across multiple deployed agents.
• Conducted structured prompt engineering, red-teaming, and iterative evaluation to improve agent accuracy and
reliability in high-stakes legal contexts.
• Collaborated with legal and technical stakeholders to translate domain requirements into AI system specifications
meeting regulatory standards

• Developed Python-based automation tools and REST API integrations to streamline internal data workflows.
• Built and tested LLM-based workflow automation pipelines for document processing and knowledge retrieval.
• Contributed to research and prototyping of AI-assisted applications across the product pipeline.