Kingz Kritiman

I’m a Full Stack AI Engineer focused on building scalable AI systems, agentic workflows, and distributed backend architectures. I bring a practical bias toward real-world impact by combining LLMs with RAG pipelines and cloud-native infrastructure.

As an AI Engineer Intern at Bitshort, I architected and implemented a core Multi-Agent AI System powering AarogyaDost. I designed a Reasoning Agent and Validator Agent adversarial loop to improve factual accuracy and eliminate hallucinated clinical outputs.

I engineered an end-to-end RAG architecture with intent-aware retrieval routing, document validation gates, and per-patient ChromaDB isolation. This reduced retrieval costs by approximately 60 percent while preventing cross-patient and cross-report data contamination, and I improved medical document extraction accuracy from approximately 30 percent to over 85 percent using an OCR reconciliation pipeline.

I also built and integrated eight specialized AI agents across clinical extraction, validation, biomarker analysis, recommendation generation, protocol generation, database reconciliation, and dashboard orchestration. Beyond features, I focused on reliability by resolving SSE timeout failures, duplicate workflow triggers, stale retrieval contexts, and cross-report answer bleed—so clinical data processing stays stable and accurate at scale.