Susan Potter

### Quant Developer & Systematic Options Trader

Production-gradequant engineering

Quantitative infrastructure for systematic trading strategies. I build the tools that make edge repeatable: backtesting pipelines, strategy validation frameworks, and the type-driven code that holds them together.

25 years engineering software systems: from trading and risk platforms at Citadel, Bank of America, Morgan Stanley, and BNP Paribas, through building performance and analytics Software-as-a-Service (SaaS) products at Northern Trust, Salesforce, and Jive. Now I am back to quantitative finance with algorithmic trading systems at Referential Labs, building strategy robustness validation, backtesting infrastructure, and portfolio risk analysis.


###

What I Do

01

Quant Infrastructure

Backtesting pipelines, strategy validation frameworks, and data infrastructure built for correctness first.

  • QuestDB time-series storage
  • Robustness testing pipelines
  • Statistical validation
02

Strategy Validation

Quantitative, statistical, and econometric methods for validating algorithmic trading strategies.

  • Backtesting with zipline
  • pandas / polars pipelines
  • scipy statistical analysis
03 λ

Type-Driven Engineering

Large codebases that survive contact with production. FP correctness applied to financial systems.

  • Scala 3 / ZIO2 pipelines
  • Haskell domain modeling
  • Property-based testing
04

Distributed Systems

Trading and risk platforms, SaaS backends, and cloud-native infrastructure at scale.

  • AWS / Kubernetes / NixOS
  • High-availability systems
  • SRE & observability

###

Current Focus

Backtesting & Validation Pipeline

Building infrastructure for validating algorithmic trading strategies using quantitative, statistical, and econometric methods. Robustness testing with QuestDB, Python, numpy, pandas, and zipline-reloaded.

Strategy Research

Applying quantitative methods to systematic strategy development. Using Pandas TA, scipy, polars, and quantdsl for signal generation and strategy validation.

Distributed Systems

25 years designing resilient distributed systems, from trading platforms at Citadel and Bank of America to SaaS backends at Northern Trust and Salesforce. Scala/ZIO, Haskell, and functional programming at scale.

Content & Writing

Author of the Git chapter in The Architecture of Open-Source Applications. Writing on functional programming, distributed systems, and quantitative methods.


###

Writing

2026-06
Quant Finding Signal in Market Noise: I Stopped Reading the Tape and Started Measuring It

How I turned visual order flow observations into testable hypotheses with computable features, from book imbalance and absorption detection to iceberg signatures and stop cascade mechanics.

2026-06
Quant A Taxonomy of Backtest Lies: Survival Bias, Lookahead Bias, and the Rest

Every backtest is biased. A catalog of the biases that corrupt results, with detection code and magnitude estimates from the literature.

2026-05
Quant From Hypothesis to Production: A Quant's Productivity Toolkit

The full validation funnel: exploration, statistical testing, Monte Carlo simulation, and robustness checks. Most hypotheses die in Stage 1, and that's the point.

2026-06
Quant Property-Based Testing Meets Financial Data

Encoding financial invariants as executable test specifications. Bid-ask spreads, OHLC consistency, monotonic timestamps: the properties that catch data problems before they corrupt your models.

2026-05
Software Event Sourcing for Financial Systems

Why storing what happened beats storing what is. Audit-ready infrastructure with Scala and ZIO, and when to skip event sourcing entirely.

2026-06
Quant Autocorrelation and What It Means for Your Backtest P&L

Autocorrelated returns inflate Sharpe ratios by 30% or more. How to detect it, correct for it, and report performance honestly.


  • "Essentially, all models are wrong, but some are useful."

    G. E. P. Box
  • "Models are to be used, not believed."

    Often attributed to John W. Tukey
  • "The map is not the territory."

    Alfred Korzybski
  • "If you torture the data long enough, it will confess to anything."

    Ronald Coase
  • "Prediction is very difficult, especially about the future."

    Often attributed to Niels Bohr or Yogi Berra

###

Technical Stack

Languages Python Scala Haskell TypeScript Erlang
Quant QuestDB pandas numpy polars zipline
FP ZIO 2 Cats Effect PureScript Property Tests
Infra AWS Kubernetes NixOS Terraform