爲岡 啓

Ubie株式会社 AI・機械学習エンジニア

自己紹介

Software Engineer at Ubie, Inc., a health-tech company building AI-driven medical solutions. I design and implement production-grade generative AI systems, focused on Retrieval-Augmented Generation (RAG), hybrid search architectures, and LLM evaluations. Currently I build automated evaluation pipelines that quantify medical safety, quality, and helpfulness, and apply Loop Engineering — iterative empirical experiment loops — to validate logic and prompt improvements before adoption. My background combines backend engineering, ML systems, and scalable data architecture. Before Ubie, I led a data engineering team at GLOBIS where I built and operated the data platform on GCP and shipped ML-driven content recommendations. At Kakaku.com (Tabelog) I worked across web apps, on-premise ML platforms (Rancher + Kubernetes), and data infrastructure. I hold a Master's degree in Computer Science from The University of Osaka. I'm interested in AI infrastructure and building reliable, real-world applications of generative AI in collaborative engineering environments.

職歴・学歴

Ubie株式会社

Technology Platform

Software Engineer, Generative AI

AI・機械学習エンジニア

Building automated medical safety evaluation infrastructure and RAG systems for Ubie's consumer-facing health AI chatbot. • Built an automated evaluation pipeline running full-dataset evals via LLM-as-Judge with context-aware criteria. • Designed eval metrics.

スキル・経験

  • データ分析
  • Ruby on Rails
  • Python
  • AWS
  • Docker
  • Kubernetes
  • Rancher
  • Machine Learning
  • GCP
  • Shell script
  • Airflow
  • BigQuery
  • Embedding search
  • Elasticsearch
  • LLM Evaluation
  • RAGAS
  • AlloyDB

言語

  • 英語

資格

  • Google Cloud Professional Data Engineer
  • TOEIC Listening & Reading Test Score: 755