About
Bridge SE with AI integration expertise, combining 4 years of software development experience with 2 years of hands-on AI/ML specialization. Proficient in PyTorch, RAG, and model fine-tuning for production deployment. Proven track record managing offshore development teams (21 members) while implementing end-to-end AI solutions.
Work Experience
FPT Japan Holdings CorporationMedical Device
Bridge SE / Systems Engineer
- Voice AI System with Local RAG: Developed on-premise dual-mode AI system for X-Ray medical device with voice-controlled operations and intelligent Q&A assistant. Fine-tuned Whisper model for Japanese medical terminology and implemented RAG architecture with LlamaIndex and Qwen 2.5-3B.
- Medical Device X-Ray Diagnostic System: Managing offshore team of 21 developers for X-Ray diagnostic workstation GUI development. Responsible for requirements gathering, quality management, code review, and technical documentation in Japanese.
Seikyo Co., Ltd.MLSystem Development
Systems Engineer
- ML-Powered Electrical Panel Quotation System: Developed Random Forest regression model for instant price prediction of complex electrical distribution panels. Achieved ±15-18% accuracy, improving sales conversion rate by 22% and reducing engineering workload by 35%.
- Quotation Management System: Built comprehensive WPF application with rule-based automation for routine quotations. Reduced quotation creation time by 40% through templates, processing 30-50 routine quotations monthly with 95% accuracy.
- Monthly Management System: Created financial management application for comprehensive business accounting, replacing manual Excel-based processes with centralized data management and automated reporting.
Projects
Voice AI System with Local RAG
On-premise dual-mode AI system for X-Ray medical device: voice-controlled operations and intelligent Q&A assistant. Fine-tuned Whisper for Japanese medical terminology with 3x inference speedup. Implemented RAG architecture with fully on-premise deployment.
Medical Device X-Ray Diagnostic System
Large-scale medical device project managing offshore team of 21 developers for X-Ray diagnostic workstation GUI development. Handled requirements gathering, basic design, quality management, and technical documentation in Japanese.
ML-Powered Electrical Panel Quotation System
ML-powered price prediction system for complex electrical distribution panels enabling instant preliminary estimates. Achieved ±15-18% accuracy using Random Forest on 2400 historical quotations. Improved sales conversion rate by 22% and reduced engineering workload by 35%.
Monthly Management System
Financial management application for comprehensive business accounting, replacing manual Excel-based processes. Developed three core modules: Cost, Sales, and Revenue management with automated monthly financial reports.
Quotation Management System
WPF-based comprehensive quotation management application replacing Excel-based workflows. Implemented rule-based automation engine for routine quotations achieving 95% accuracy. Reduced quotation creation time by 40% through templates, processing 30-50 quotations monthly.
Personal Projects
IoT-Based Safety Helmet Detection System
Graduate project: IoT edge computing system for real-time safety monitoring. Fine-tuned Tiny YOLO on Raspberry Pi 3 for helmet detection achieving 87% accuracy. Created custom C# WinForms annotation tool, completing 2500 images in 2 weeks.
news.hungba.net
AI-powered news aggregation with OpenAI API integration for automated summarization.
jlpt.hungba.net
JLPT (N1, N2, N3) practice platform with spaced repetition system and AI-generated grammar explanations.