Aki Ariga
Aki Ariga

Staff Software Engineer

AI Product Engineer
Download CV
Hello, I’m Aki👋
As a Vancouver-based Staff Software Engineer at LayerX, I specialize in Human-in-the-loop systems, ML product development, and MLOps. I am driven by the mission to use machine learning and AI to make a meaningful difference for both businesses and society.
Recent Posts

Embedding workflow templates in skills: shifting the LLM's role from "generation" to "rendering"

The dream of LLM-powered ML workflow generation At my company, an ML feature called AI Signals provides capabilities like RFM analysis, recommendation, and contextual bandits. To …

Migrated from Pages CMS to Sveltia CMS featured image

Migrated from Pages CMS to Sveltia CMS

I introduced Pages CMS last summer, but after encountering several concerns, I migrated to Sveltia CMS. Motivation As I wrote in this comment on GitHub, the move was triggered by …

Between Principal and Glue Work

Introduction I have always worked with the mindset that being a Staff+ Engineer is essentially doing high-level “glue work.” Because the tasks are surprisingly diverse, I have …

Tackling Review Fatigue by Document Driven Agentic Coding featured image

Tackling Review Fatigue by Document Driven Agentic Coding

This year, the amount of time I spend on reviews has exploded. This applies to both code and documentation. And the fatigue from this has also increased dramatically. Of course, …

Configured Pages CMS featured image

Configured Pages CMS

I had been looking for a way to write Hugo articles on mobile devices like my iPad, and that’s when I came across this article by mehori and decided to try Pages CMS.

Recent Posts (Japanese)

AI時代の転職活動記録

今回、転職活動をするにあたって、AIをフル活用して活動をした。前職ではGithub Copilotが個人アカウントだが会社にコントロールされていたので、せっかくだからとClaude Proを契約して望んだ。

LayerXで働き始めました

友人である riywoさんに誘われて、4/1にLayerXのAi Workforce事業部に入社しました。居住地は引き続きバンクーバー(の2つ隣の市です。いわゆる「神奈川」の代表点としての「横浜」みたいな説明)です。

Treasure Dataを退職しました featured image

Treasure Dataを退職しました

2026/3/31付けで7年半務めたTreasure Dataを退職します。

ワークフローテンプレートをskillに埋め込んで、LLMの役割を「生成」から「レンダリング」に変えた話

MLワークフローのLLMによる自動生成の夢 会社のAI SignalsというML機能では、(MLではないが)RFMや推薦、Contextual Banditといった処理を行う機能がある。特にMLの予測処理をスケーラブルに実行するために、ML用APIをcallすると裏側でAWS Batchが起動して並列でworkerを立ち上げるという仕組みを取っている。この並 …

Montréalに10年以上ぶりに行った featured image

Montréalに10年以上ぶりに行った

3月の上旬にworkationとして友人を訪ねがてらMontréalとOttawaに数日行ってきた。

Featured Publications
MLOpsの歩き方 (Beginners Guide to MLOps) featured image

MLOpsの歩き方 (Beginners Guide to MLOps)

This article covers very biginning guide for MLOps, i.e., What is MLOps? How do tech giants make Machine Learning systems? What challenges are important? or major open sources for …

avatar
Aki Ariga
仕事ではじめる機械学習 (Machine Learning at Work) featured image

仕事ではじめる機械学習 (Machine Learning at Work)

First book for how to design Machine Learning systems and how to proceed Machine Learning projects. This book is originally written in Japanese, but available in Korean, Simplified …

avatar
Aki Ariga