Toru Hosoi — career history of a Cloud Infrastructure Engineer focused on AWS design, build, and operations for large-scale video streaming platforms.
For about eight years I've run infrastructure for a Japanese video streaming platform. With millions of users, most of my time goes into designing and preparing for traffic spikes so the platform stays up when everyone arrives at once.
All infrastructure is managed with Terraform, with no hand-rolled resources. CI/CD is set up alongside, so anyone can reproduce the same environment.
Designed and built the AWS environment as the dedicated infrastructure engineer. Terraform-based IaC, private network design via DirectConnect, and container-based architecture on ECS/RDS. Stack: DirectConnect, CloudFront, ALB, RDS, ECS. Tools: Terraform, GitHub.
Requirements, infrastructure design, build, and operations. CDN design on CloudFront with WAF for security. CI/CD pipelines on CodePipeline. Pre-scaling and auto-scaling tuned for traffic spikes driven by program schedules. Event-driven monitoring and notifications via PagerDuty. Stack: CloudFront, WAF, S3, ALB, EC2, ECS Fargate, Lambda, Aurora MySQL/PostgreSQL, ElastiCache, OpenSearch, DirectConnect, CloudWatch, EventBridge, CodePipeline. Tools: Terraform, GitHub, PagerDuty, Docker. OS: Amazon Linux family.
Database design and build support for a large-scale user platform. High-availability and high-scalability architecture using Aurora MySQL and DynamoDB.
Built a real-time streaming platform using AWS video-call features. End-to-end infrastructure build and operations, tuning and monitoring for stable video delivery. Stack: CloudFront, ALB, ECS, RDS, Chime.
Designed and built infrastructure for a web application. Standard web architecture on CloudFront, ALB, EC2, and RDS. SSL certificate management and network configuration.
Field testing of Wi-Fi quality at retail and convenience store venues. Data preparation for monitoring terminals and supporting administrative work.
Improvement proposals, log analysis for user-reported connectivity issues, field testing (GPS and call quality verification), visualisation of results, and recommendations for specific improvements.
Software upgrades and incident response for base stations. Coordinated on-site technicians and performed remote state checks and configuration work.
Pre-release software verification in an anechoic chamber. Deployed and validated development builds.
Hardware replacements, new software rollouts, and incident response. Directed on-site technicians, performed remote configuration, tracked team progress, and handled customer escalations.
Kitting 5,000 PCs and software testing. Built master images and deployed them at scale. Stack: Windows, OS kitting tools.
Designed, built, and rolled out DNS, mail, and proxy servers. Configured L2/L3 network switches. Stack: Linux, UNIX, AIX, Solaris.
I've worked on high-impact services — video streaming, public-sector systems — where an outage really matters. Load tests, pre-scaling, and CloudFront cache strategy together keep the platform steady during traffic spikes.
Terraform keeps environments reproducible and cuts manual work. CodePipeline-based CI/CD gives safe, fast release flows without sacrificing guardrails.
Base-station operations and field work at a telecom carrier gave me a feel for physical layer and operational constraints. I bring that awareness into cloud architecture — "what actually happens in the field" shapes the design.
During a Slack workspace migration, I used Cursor + Claude to build a Lambda-based Slack bot that kept posts in sync before and after the cutover. It eliminated the data-inconsistency problem and made the migration clean.
My handle is thorhosoi across platforms.