SUHAAS VIJJAGIRI

Architected a framework-agnostic evaluation system for enterprise RAG pipelines, ensuring faithfulness and precision. Orchestrated synthetic data generation using Knowledge Graphs to enable robust adversarial testing of LLM applications in a regulated healthcare environment.

Led the end-to-end engineering of 'KingbotGPT', a production RAG chatbot serving 30k+ students. Designed advanced 'LLM-as-a-Judge' evaluation frameworks and integrated full-stack observability (Langfuse) to optimize response quality and minimize hallucinations.

Built scalable, real-time analytics platforms for aviation IoT. Developed complex front-end interfaces and optimized backend performance for high-frequency data visualization, bridging the gap between physical asset tracking and digital decision-making.