↑ Promoted internally after 4 months

• Sole technical decision-maker across LLM provider selection, model deployment strategy, agent graph design, retrieval pipelines, and production-readiness criteria for 3+ simultaneous AI systems.
• Designed a multi-agent software delivery system covering 6+ specialized roles (product ownership, architecture, DevOps, QA, security, development) — ~90% accuracy in human-reviewed code generation outputs.
• Designed a production tender intelligence system ingesting 200+ procurement opportunities/day, generating qualification scores, summaries, and proposal drafts in under 30 minutes.
• Mentor and code reviewer for ~6 junior engineers, establishing architecture standards, security processes, and production-readiness criteria.

• Contributing to applied research on multi-agent LLM systems, RAG pipelines, and evaluation frameworks for clinical decision support in liver disease and oncology.
• Co-designed multi-agent HCC clinical reasoning architecture leading to ASCO 2026 abstract acceptance.
• Designed RAG-augmented LLM pipeline for MASLD clinical QA; accepted as THU-212 poster at EASL 2026.