Skip to content
Roy Yang

Senior data scientist, banking and insurance

Roy Yang

Over 15 years in banking and insurance, more than 9 of them applying ML forecast and analytics to liquidity, risk and profitability. On my own time I build the whole pipeline: data, model, and the system that runs it.

In banking and insurance
15+ yrs
In banking and insurance
In analytics and AI
9+ yrs
In analytics and AI
Quant and applied AI research
Point-in-time
Quant and applied AI research
View projectsRésuméContact

The pattern behind my work

Models

Prices, macro, filings, news

Market data

15-minute bars for 75 symbols since 2010, FRED macro series with their revision history, SEC filings, news and prediction markets. Ten collectors write to PostgreSQL, and every fetch is logged with what it got and what failed.

  • PostgreSQL
  • IBKR API
  • FRED
See the project
The pattern behind the projects below: data in, models and agents propose, deterministic code decides. Select a stage to see where it's built.

01Selected work

Systems I've built

Personal research and engineering projects, from point-in-time research to AI agents and the engine that runs strategies. Each one lists the skills it shows.

Recorded run, step through it

AI trading swarm

A committee of AI analysts that reads point-in-time market data, cites every claim and hands its ideas to Strategy Orchestra to test, with a human approving each idea and every number reproduced before a strategy counts.

Watch one real idea go through the swarm: five analysts and a Critic on TQQQ, the thesis I approved, and the campaign that tested it, from 9,826 trades stopped at a gate to the Judge's 2.4 out of 100. Every line comes from the run's own records.

  • Python
  • PostgreSQL
  • TradingAgents
  • LangGraph
  • LM Studio
  • FastAPI
  • React
Read the write-up
royyang.ai/demos/ai-trading-swarm
RECORDED RUN · TQQQ
  1. Data
  2. Analysts
  3. Critic
  4. Thesis
  5. Owner
  6. Orchestra
  7. Judge
  1. DATAas of 24 Sep 2026 · 7 of 10 sources ready, 3 stated missing
01Forecasting and ML

TSBench: multi-series forecasting benchmark

  • Python
  • pandas
  • statsforecast
  • LightGBM
  • arch

A benchmark harness that runs forecasting models from several libraries over 16 years of 15-minute bars for 75 US tickers, scores them against finance-specific baselines, and picks winners with a Model Confidence Set.

15-minute bars, 75 tickers, 2010–2026
7.2 M
15-minute bars, 75 tickers, 2010–2026
known-answer tests pass on production data
10 / 10
known-answer tests pass on production data
horizons where volume beats seasonal naive
5 / 5
horizons where volume beats seasonal naive
02Reinforcement learning

Reinforcement-learning trading strategy

  • Python
  • stable-baselines3
  • Gymnasium
  • PyTorch
  • TA-Lib

Two agents, PPO and A2C, choose which of 10 low-correlation ETFs to trade every 15 minutes, and how much. The pipeline runs end to end. The evaluation doesn't hold up yet, and this page shows why.

ETFs kept for low correlation (mean ρ 0.12, against 0.42)
10 of 23
ETFs kept for low correlation (mean ρ 0.12, against 0.42)
inputs the agents read at every 15-minute step
21,455
inputs the agents read at every 15-minute step
valid out-of-sample results so far
4
valid out-of-sample results so far
See all projects

02What I work on

From research notebook to running system

The problems I spend my time on, inside banks and on my own.

03Skills and tooling

The stack I reach for

Select a skill to see the projects that use it.

Quant research and ML

Skills, with the projects that show them

Tools

Python, DuckDB, turbovec, Jupyter

AI agents and LLMs

Skills, with the projects that show them

Tools

Claude Code, Ollama, LM Studio, Docker

Systems and execution

Skills, with the projects that show them

Tools

asyncio, SQLite (WAL), Interactive Brokers API, YAML config

04About me

Inside the banks, and on my own

I am a data scientist with over ten years in banking and insurance, the last several centered on analytics and machine learning. I have led research and quantitative strategy for global corporate banking, built comprehensive executive dashboards for corporate banking strategy, and developed machine learning forecasting models, liquidity analytic models, and risk analyses. My models have leveraged cashflow forecast results to drive consistent direct profit growth.

My path wasn't a straight line. I started on the front line of banking, advising clients, then moved into middle office for project and global corporate banking strategy, and along the way earned a master's in financial computing. That mix of client-facing experience, collaboration with both internal and external stakeholders, finance business knowledge, and quantitative and qualitative training shapes how I work: start from the decision someone actually needs to make, then build the model that informs it for practical outcome with direct measurable impact.

any data analytics related roles in financial services.

More about meGet in touch