Kyoko Shimada, Developer in Portland, OR, United States
Kyoko is available for hire
Hire Kyoko

Kyoko Shimada

Verified Expert  in Engineering

Database Developer

Portland, OR, United States
Toptal Member Since
April 7, 2022

With over 20 years of software development experience, Kyoko has touched every aspect of data-driven decision-making systems. She has architected and built quantitative hedge fund research and testing platforms at one of the largest financial institutions. She also has architected and built an A/B testing platform at one of the largest media companies, where they run 3,000 experiments daily with 200 metrics for 75 million active users.

Amazon Web Services (AWS), Python 3, Apache Airflow, Data Build Tool (dbt)...
D3X Systems
Python 3, Java
Pandora Media Inc.
Google Cloud Platform (GCP), Scala, Python 3, Apache Airflow




Preferred Environment

Java, Python, Linux, MacOS

The most amazing...

...platform I built was a "lineage" tool which could answer any question about the company's data such as lineage, usage, retentions, freshness, and anomaly.

Work Experience

Staff Data Enigneer

2020 - PRESENT
  • Implemented the first real-time machine learning (ML) service for the company.
  • Served as a technical lead in building the company's central core models.
  • Mentored and guided mid-level data engineers. Named as a role model by two team members.
Technologies: Amazon Web Services (AWS), Python 3, Apache Airflow, Data Build Tool (dbt), Redshift


2019 - 2020
D3X Systems
  • Built a quantitative portfolio management system that optimizes theoretical alphas into tradable assets accompanied by visual analytics of risk and returns in various dimensions.
  • Wrote a detailed white paper on efficiently optimizing portfolios using Grinold and Kahn Active Portfolio Management theory.
  • Created a Python library to analyze theoretical models.
Technologies: Python 3, Java

Staff Data Enigneer

2014 - 2019
Pandora Media Inc.
  • Served as a technical lead and guided migration from on-premises to Google Cloud Platform.
  • Led and built a visualization tool of over 40,000 on-premises Hadoop jobs.
  • Led, architected, and implemented Pandora's A/B testing platform. Analyzed behaviors of 75 million active listeners for over 3,000 experiments and 200 metrics daily.
Technologies: Google Cloud Platform (GCP), Scala, Python 3, Apache Airflow

VP & Director

2005 - 2014
  • Performed alpha research within the Scientific Active Equities Hedge Fund department. Focused on text analysis using the Java Lucene library.
  • Proposed, architected, led, and built an alpha research platform using Python, Pandas, Django, Sphinx, and D3.js.
  • Led a mid-size team of data science engineers for a multi-billion dollar quantitative macro hedge fund. The team was responsible for complex data processing, optimizing portfolios, generating alpha forecasting models, and building risk analysis tools.
Technologies: Python, Java, MATLAB, D3.js, Pandas

Senior Software Engineer

2001 - 2003
Lehman Brothers
  • Led technical projects within the Equity Finance and Prime Brokerage departments.
  • Architected and a built data warehouse for the prime broker business.
  • Architected and built various risk and portfolio summary tools.
Technologies: Java

Software Engineer

1998 - 2000
Goldman Sachs
  • Built an internal web portal for the Goldman Sachs Asset Management (GSAM) department.
  • Assisted on Y2K compliance projects, which required many testing and documentations required by SEC regulations.
  • Received a three-month intensive training on financial engineering and business etiquettes.
Technologies: Java


While leading a massive migration effort from on-premises Hadoop systems to GCP, I felt the need for a tool to analyze tables and jobs efficiently and track the migration status. I started implementing a tool to answer all sorts of questions about our data, and the tool called Lineage was born. It was a Django-based application where I scraped various logs and fed the data to the UI. The users were able to add their own comments and descriptions. The tool showed every data usage, freshness, anomaly, retentions, lineage, similar tables, and other tables often used in conjunctions.


Being frustrated with the lack of efficient tools while developing new quantitative financial trading models, I initiated and developed a quantitative hedge fund research tool, Alpyne. It's a Python-based modularized platform that lets you easily implement new ideas, perform historical backtests, and provides numerous risk and return visual analytics. The tool gained so much recognition that I still get inquiries from various hedge funds to build similar tools for them.


Java, Python, Scala, Python 3


Apache Airflow, MATLAB


Amazon Web Services (AWS), Google Cloud Platform (GCP), Jupyter Notebook, Pick D3


Redshift, Apache Hive


Data Build Tool (dbt), Applied Mathematics


Django, Hadoop


D3.js, Pandas

2000 - 2004

Master's Degree in Mathematics

New York University - New York, New York, USA