-
I've spent three years as a web engineer and seven years as a research engineer at an online advertising company. My research interests lie in computer vision, natural language processing, and machine learning. The immediate mission is to scale R&D in my company via engineering such as MLOps and ResearchOps.
-
-
2017 - Present
Research Engineering (AI Lab)
Scale R&D in my company via engineering such as MLOps and ResearchOps
NLP, Computer Vision, Python, TensorFlow, PyTorch, TypeScript, Kubernetes, Terraform, GCP, AWS2016 - 2017
Backend Engineering and Data Science (iXamDrive)
An operation platform for in-feed advertising. It can publish dynamic ads social media such as Facebook, Twitter, and LINE as official partner.
Scala, Java, Finagle, Python, TensorFlow, AWS, GCP2015 - 2016
Backend Engineering and Front Engineering (Game Tailor)
A dynamic retargeting ad product for smartphone apps. Tailor-made ads generated based on each game user status such as game level, selected characters and log in frequency will be served to worldwide game users through social medias.
Scala, MySQL, Aerospike, Apache Spark, RedShift, AWS2014 - 2015
Backend Engineering and Front Engineering (GameLogic/Audience, Smalgo)
An online advertising system as demand-side platform (DSP) to focus on mobile game publishers.
Java, MySQL, Redis, Ruby on Rails, Angular, AWS -
17-Apr, 2025
Skill Up Boot Camp for Research Scientists
Skill up onboarding for researchers were held and materials are available.
25-Dec, 2024
User Embeddings from their skill set
Attempts to reveal people skill sets based on their text (e.g., slack chats, source codes, papers, etc.).
13-Oct, 2021
Easy-to-use experiment management hands-on with MLflow in GKE
Experimental management systems using kubernetes clusters for researchers.
1-Jun, 2020
MLOps2020: Start Small, Grow Big
The effective use of MLOps tools such as Hydra, Sacred, MLflow, Kedro, Metaflow, Kubeflow, and Optuna is explained with a comparison of their features.
5-Feb, 2019
Tweet-based matching system technologies
An installation work that extracts user features from their tweet and pre-questionnaires to realize a dialogue in an Augmented Reality space.