Yuki Mabuchi
Backend Engineer @ en Inc.
About
I'm a backend-focused software engineer working across infrastructure, recommendation systems, and MLOps. Experienced in designing and operating large-scale recommendation platforms for job marketplace services with 300K+ MAU, integrating search engines, machine learning models, and AWS infrastructure. I love video games and creating something. I create mobile games in my spare time! My name is 馬渕勇貴 in Kanji. Personal Github profile (https://github.com/YukiMabuchi) Zenn tech blog (https://zenn.dev/mabuchi1118) Current tech stacks Languages: Python Frameworks: FastAPI DB: OpenSearch, Redis, MySQL, DynamoDB Infra: AWS, Terraform Others: Docker, Git, GItHub Actions, New Relic
Japan
Tondabayashi
Computer Software
Docker製品, Terraform (ソフトウェア), OpenSearch, Redis, TypeScript, PostgreSQL, Unity, C#, JavaScript, PHP, Lua, MySQL, postgress, アマゾンウェブサービス (AWS), Python (プログラミング言語), Next.js, 英語, React Native, TOEIC, IELTS
Experience

Backend Engineer
- Leading an MLOps Initiative involving around 10 members and EMs - Defined the organization’s MLOps model and aligned both teams on ownership and operating principles. - Restructured team boundaries and responsibilities to address duplicated ML code and validation overhead caused by inconsistent data sources. - Leading a shared library development for feature preprocessing and embedding generation, reducing model integration and validation lead time by about one month. - Mentored the AI team on engineering practices such as design, PR reviews, and development workflows, helping standardize the software development process. - Led an end-to-end ranking model replacement for a large-scale recommendation system, covering requirements definition, system design, implementation, validation, and production release planning. - Designed an extensible ranking architecture using Clean Architecture, enabling safe future model swaps with minimal blast radius. - Coordinated development and release schedules across ML and backend teams, and introduced load testing into pre-release checklists to ensure readiness for peak traffic. - Identified and resolved TensorFlow Serving latency bottlenecks introduced by a new model by applying concurrency/parallelism and tuning CPU, memory allocation, and ECS task scaling, which improved average response time from 500ms to 100ms. - Optimized ETL jobs processing 1–2 million records per run by introducing chunked processing and Python generators, reducing average ECS batch container memory usage by 50%+. - Reduced API latency by refactoring sequential DynamoDB BatchGetItem calls into concurrent execution using Python’s ThreadPoolExecutor (200ms → 100ms). - Improved engineering productivity by standardizing AWS resource naming across Terraform modules and introducing PR size guidelines and PR metrics tracking, reducing average PR close time from 2 days to 1 day.

Full-stack Developer
新横浜
Experiences 1. Agent AI Project: (November 2023 - Present) Development of an office tool like your real assistant. This project is one of the two new products in our company and I and the other members started developing from scratch, using Next.js and PostgreSQL as our first attempt in the company. I am mainly in charge of database design and admin console development of both frontend and backend. Main Responsibilities: - UI design, database design, frontend, and backend implementation for the development of our application’s admin console. - Integration with external services such as Slack, Teams and Lark to strengthen service demand. - Implementing features using OpenAI’s Assistant API to obtain responses (Partly). - Setting up development servers using Linux and AWS (Partly). Tech stack: Next.js, TypeScript, Python, PostgreSQL, AWS, Linux. 2. LLM Project (April 2023 - November 2023) Development of chatbot functionality that is more free-form and closer to natural conversation than template responses. Main Responsibilities: - Development of features on the admin console including Email notification system, error log management system, unknown chat response list and so forth. - Implementation of a web search feature using APIs provided by Google and Microsoft. - Integration of OpenAI API with our own chat system (Partly). Tech stack: JavaScript, PHP, Lua, Python, MySQL, AWS, Linux, LangChain 3. Game Campaign Project (September 2022 - November 2022) Development of a feature to distribute prizes through campaigns using animations such as gacha as a feature for mini-application operating on LINE and the web. Main Responsibilities: - Implementing the landing page. - Implementing features like automation test, report and version management systems for safer and more efficient service deployment. Tech stack: JavaScript, PHP, Lua, MySQL, AWS, Linux 4. Bug fixes, Styling (July 2022 - September 2022)

事務職
リフォーム産業新聞社
日本 東京都
I have been engaging in social network service related tasks, proofreading pre-publication articles, and researching social issues that may affect reuse field.
Yuki Mabuchi's Contact Information
Phone
Find the Right Leads
Find Verified Contact Data
What LeadContact does well
Find verified emails, phone numbers, and decision-makers with 98% accuracy.
Find Leads
Find the right people by company, role, industry, location, and more.
925M+ professional profiles

Find Emails
Access verified email addresses for your target contacts.
657M+ emails

Find Phone Numbers
Get cross-validated phone data from multiple top sources.
239M+ phone numbers

More Accurate. Lower Cost.
Find contact data in 1 tool with 98% accuracy
LeadContact integrates leading enrichment tools to deliver more accurate contact data—without paying for each one.
Great conversations start with the right contact.
It’s time to find yours.


