Enterprise data, analytics, and AI workflow initiatives.
01
Enterprise Data Integration Platform
Designed and led an enterprise data integration platform connecting 20+ source systems across CRM, MIS, school systems, and HR platforms. The platform supported 12+ business units across 5+ countries and enabled reliable data movement for analytics, reporting, and operational decision-making.
Led the modernization of business intelligence and decision-support reporting across Sales, Marketing, Operations, HR, Academic, IT, Finance, and Materials Management domains. The initiative transformed fragmented reporting processes into standardized dashboards, semantic models, and executive-ready insights.
Designed AI-assisted workflow automation for insight summarization and document analysis, helping business teams reduce repetitive analytical work while keeping human review in the decision loop. The solution combined Python, APIs, LLM capabilities, automation scripts, and n8n-based workflow orchestration.
10 → 2manual steps reduced to guided steps
4h → 10mtask completion time reduced
100+reports, documents, or requests processed per day
Designed and built a retrieval-augmented AI assistant that answers "how do I…?" questions about iSAMS school management software, grounded in a curated knowledge base rather than general model knowledge. Shipped end to end as a production system with streaming chat, an admin console, and a feedback loop that turns approved answers back into knowledge-base improvements.
Built a personalised tour-advisory chatbot for the Vietnamese market that holds context across a conversation instead of treating each message as a fresh search. Conversation state drives hard SQL filtering before vector search, and a controlled closing flow hands a qualified lead to human sales rather than pretending to complete a booking.
96backend tests covering unit and integration paths
2-layerNLU — deterministic rules first, LLM only when ambiguous
4-stateclosing flow from exploring to a qualified lead