Innovative AI Solutions &
Technical Leadership

Architected a secure framework coordinating 20+ specialized LLM agents. Automates the full software development lifecycle via specialized agents (Agentic Team comprising BA, Dev, QA/QC, and Security), dependency checks, and synthetic data verification.

Designed and deployed a scalable multi-tenant SaaS Chatbot platform with modular plugin support and Omni-channel integration. Serves both Internal corporate chatbots and commercial SaaS setups, using advanced RAG and a MAP Portal for SaaS management.

Built a production hybrid semantic-lexical search pipeline over 5 million technical files, leveraging semantic chunking and reranking models. Optimized search speed by 45%.

Intelligent optical character recognition (OCR) system for automatic port gate control. Utilizes OpenCV, YOLOv8, and TensorFlow for real-time license plate and container code recognition with high accuracy.

Highly secure electronic Know Your Customer (eKYC) platform featuring passive facial liveness detection, face matching, and automated ID card parsing with deep learning.

Engineered a high-performance inventory microservice architecture handling millions of concurrent users built with .NET 10 templates for high reusability and rapid deployment. Implemented gRPC, Kafka queues for event-driven messaging, and Hybrid Caching (Memory + Redis).

Built a high-throughput, low-latency Big Data platform using CDC (Change Data Capture) to stream and process enterprise data in real-time. Integrated an AI-powered Observability Agent to automatically analyze multi-source logs, detect anomalies, and trigger instant system alerts.

Engineered a low-latency API gateway managing massive incoming prompts, integrating advanced Web Application Firewall (WAF) protection, sliding-window rate limiting, dynamic fallback routing, and hybrid caching layers.

A structured output definition template for Claude LLM, optimized for building independent deterministic AI agents.

A production-ready skeleton codebase following DDD/CQRS patterns, integrated with Native AOT for high-speed microservices development.

Core contributor and architecture designer for Loop Engineering, an open-source methodology by Cobus Greyling for running autonomous AI coding agents (Claude Code, Cursor, Grok) reliably in production, since May 2026.

Internal enterprise chatbot platform for Saigon Newport Corporation (SNP), covering automation and chatbot design for internal operations.

Designed and deployed an automated, AI-driven code review workflow integrated into the CI pipeline, using MCP to connect with document repositories and source control.

A real-time vision assistant that describes surroundings aloud to help visually impaired users navigate their environment.

A production-grade reference architecture for enterprise knowledge management, combining semantic search with a Knowledge Graph so AI agent retrieval stays within corporate governance and access-control boundaries.

A security-focused automation platform using multi-agent architectures to proactively scan, detect, and mitigate vulnerabilities such as SQL injection, hardcoded secrets, and missing input validation.

A multi-agent team framework coordinating specialized AI agents to work together on shared engineering tasks.

An internal MCP portal managed through a SASE security layer, centralizing access to MCP servers and Claude Skills.

An open-source reference architecture for building a production-grade enterprise Retrieval-Augmented Generation (RAG) and Knowledge Graph (KG) engine.

A self-learning system bridging AI agents and a team's accumulated knowledge base, creating a persistent experience layer that lets AI adapt to specific workflows.