Skip to content
View heywoodwt's full-sized avatar

Highlights

  • Pro

Block or report heywoodwt

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
heywoodwt/README.md

Heywood Williams-Tracy

M.S. in Data Science, University of Virginia School of Data Science (2026). B.A. in Public Policy & Data Science, UNC Chapel Hill (2023).

Instructional Assistant for DS 2002: Data Science Systems at UVA — the course covering the engineering layer beneath data science work: relational and NoSQL data stores, ETL pipelines, cloud services, containers, and API integration. I held office hours, walked students through lab environments, and debugged their pipelines with them. Explaining why a broken pipeline is broken, to someone who has not yet learned to read the error, turned out to be the fastest way to get rigorous about my own fundamentals.

Based in Charlottesville, VA. Founder of warppool. Open to data science and ML engineering roles.


What I work on

Quantitative trading systems. Market-making bots for Kalshi's prediction markets — live order management, inventory and risk limits, and backtesting against recorded order book data. The current generation is a Rust rewrite of an earlier Python system, built as a deliberate line-for-line port so the two could be diffed against each other under golden-fixture parity tests.

Data engineering. Serverless ETL and alerting pipelines on AWS — Lambda, EventBridge, DynamoDB, SNS — defined as infrastructure-as-code with SAM and Terraform. Also Docker, Kubernetes, and Kinesis, from both DS 2002 and UVA's cloud systems coursework.

Applied statistics and ML. Regression and statistical learning, deep learning (CNNs, sequence models), computer vision, and geospatial analysis in Python and R.


Selected public work

Project What it is
kalshi-mm Rust market-making bot for Kalshi prediction markets. A line-for-line port of an earlier Python system, with golden-fixture parity tests proving the two agree on fees, observations, and policy output. 7k lines, 136 tests, CI, containerized release.
job_alert_bot Serverless AWS pipeline that scrapes job listings, deduplicates against DynamoDB, and alerts via SNS. SAM IaC, 31 tests under moto.
Cville_Parking_Ticket_EDA Geocoding and interactive mapping of Charlottesville parking enforcement data.
PGB_FantasyFootball Season-long analysis off the ESPN Fantasy API; reconstructs PPR scoring from raw stat lines.
Flood_Analysis_PLCY_698 County-level flood risk and disaster-relief funding analysis in R, for the Environmental Defense Fund.
ds_5030_work Graduate statistical learning coursework — regression, classification, resampling, regularization.

Some of my strongest work like my MSDS capstone is still in private repositories. Happy to walk through any of it.


Tools

Python · Rust · R · SQL · pandas · scikit-learn · PyTorch · AWS (Lambda, DynamoDB, SNS, Kinesis, SAM) · Docker · Kubernetes · Terraform · Git


Website · LinkedIn

Pinned Loading

  1. kalshi_v5 kalshi_v5 Public

    Kalshi BTC market-making bot: WebSocket ingest, RL policy, and the Python reference implementation behind the Rust port in kalshi-mm

    Jupyter Notebook

  2. kalshi-mm kalshi-mm Public

    Rust live market-making bot for Kalshi (behavior-parity port of the Python live_trader_v2)

    Rust

  3. job_alert_bot job_alert_bot Public

    Serverless AWS pipeline that scrapes job listings, filters them by keyword, and sends SNS alerts. Lambda + EventBridge + DynamoDB, deployed with SAM.

    Python