Lekha Vulavala

• Designed and evaluated diverse prompts to identify hallucinations, inconsistencies, and biases in generative AI models.
• Collaborated with ML engineers to improve model safety and factual accuracy using systematic evaluation frameworks.
• Leveraged LLM evaluation strategies to detect safety violations, toxic language, and prompt injection vulnerabilities.
• Performed model performance monitoring using key metrics (F1-score, precision, recall) across different prompt structures.
• Contributed to internal dashboards tracking evaluation scores and model behavior patterns.

Key Skills: Prompt Engineering, LLM Monitoring, Bias Detection, Model Evaluation, Generative AI

• Built and monitored AI solutions using OpenAI and Azure-based services.
• Worked with large datasets to validate models, analyzing model drift and performance degradation over time.
• Participated in fairness testing by evaluating model outputs across demographic segments.
• Developed dashboards and visualizations using Power BI and Python for stakeholders.

Key Skills: Azure AI, Model Monitoring, Bias Evaluation, Visualization, Cloud ML Pipelines

• Developed and maintained responsive and user-friendly web applications using HTML, CSS, and JavaScript, driving an improvement in user engagement metrics.
• Collaborated with cross-functional teams to understand requirements, translating them into intuitive UI/UX designs that increased customer satisfaction scores.
• Leveraged Angular to build interactive web interfaces, reducing time-to-market for new features.
• Executed RESTful APIs to integrate with backend services and databases seamlessly.