Bio

Hi, I’m Yoshihiro Hori, a ML engineer at SyntheticGestalt, where I read papers, build models, manage data pipelines, and set up and maintain infrastructure.

I’ve worked with a variety of data that are used to build predictive models, including images, videos, protein sequences, and molecular structures, ranging from thousands to tens of billions of records.

As I spend more time as a practitioner, I’ve become increasingly fascinated by the software ecosystem that makes it all possible. As such, I’m contributing to conda-forge by night and also fiddling around with Nix in my spare time.

Technical Skills

Infrastructure & Orchestration

  • Terraform (primarily for building k8s clusters)
  • Kubernetes

Data Engineering

  • Data Orchestration
    • Dagster
  • Data Transformation
    • Ray (Ray Data)
  • Data Wrangling
    • PySpark/DuckDB/Polars etc…

Machine Learning

  • Developing Models
    • PyTorch
  • Serving Models
    • Ray (Ray Serve)
  • Selling Models
    • Amazon SageMaker

Hobbies

  • Riding bicycles
  • Climbing boulders
  • Brewing coffee

Contact