Tagy Aldeen

I'm a backend engineer who has spent the last eight years designing and running distributed systems in production — and the last few also leading the teams that build them. At Monstarlab in Tokyo I'm a Senior Lead Software Engineer: technical authority on enterprise delivery, from architecture and authentication design to estimates for client bids, and mentor to engineers across several project teams.

The systems I've owned include a Japanese FinTech platform serving 40 million users (leading its Java-to-Kotlin migration while holding p95 latency around 160 ms), a multi-tenant B2B logistics platform on AWS with multi-region DR, and GCP microservices for a marketplace serving millions. I care about the unglamorous parts: data consistency across services, observability, incident response, and turning person-dependent knowledge into designs and runbooks.

More recently I've been building LLM systems for real workloads — RAG pipelines on AWS Bedrock and OpenSearch, and multi-agent orchestration for sales automation. Before industry I did a Master's in deep learning at Nagoya Institute of Technology on a MEXT scholarship, publishing on ML-based intrusion detection for IoT.

I write here about backend architecture, running systems in production, and what actually works when you put LLMs behind an API.