Experience

Lead backend engineering for a Japanese SaaS company, managing architecture decisions, team growth, and production systems serving thousands of users across Japan.

• Grew backend team from 2 to 6 engineers while maintaining high delivery pace
• Architected migration from monolith to microservices, reducing deployment time by 83%
• Led integration of GPT-4 and Gemini APIs for customer-facing AI features
• Established engineering standards: code review process, incident response, on-call rotation
• Collaborated directly with Japanese leadership on product roadmap and technical strategy

Designed and implemented cloud infrastructure on AWS for production workloads, focusing on reliability, security, and cost efficiency.

• Designed multi-AZ AWS architecture supporting 99.9% uptime SLA
• Implemented Infrastructure as Code with CloudFormation, eliminating manual provisioning
• Built CI/CD pipelines with GitHub Actions, reducing release cycle from weekly to daily
• Reduced AWS infrastructure costs by 35% through reserved instances and right-sizing
• Established security posture with IAM least-privilege, VPC isolation, and CloudTrail auditing

Built REST APIs and backend services for the core product, including payment integrations and notification systems.

• Developed RESTful APIs consumed by iOS, Android, and web clients
• Integrated Stripe and GMO payment gateways for Japan market billing
• Built push notification service using AWS SNS supporting 50K+ devices
• Improved API response time by 60% through query optimization and Redis caching

Contract research and engineering work on deep learning for medical IoT applications, resulting in a published Elsevier Q1 paper.

• Designed CNN-LSTM model for COVID-19 detection from e-nose sensor time-series
• Built data preprocessing pipeline handling sensor drift and environmental noise
• Achieved 94% classification accuracy, exceeding project targets
• Optimized TensorFlow model to TFLite for Raspberry Pi edge deployment
• Co-authored paper published in Elsevier Q1 journal on deep learning anomaly detection