Consulting

Available for Consulting

I work with teams who want to move beyond dashboards and toward decisions — using Bayesian methods, causal reasoning, and rigorous experimentation to extract real insight from messy data. With a background in formal logic and philosophy of causation (MSc Logic, ILLC Amsterdam), I bring conceptual rigour to hard applied problems in data science, analytics, and experimentation.

I consult as a Principal Data Scientist with PyMC Labs, the leading Bayesian consulting firm, where I have delivered multiple enterprise engagements across retail and technology sectors — including projects with $1M+ in measurable business impact. If your team is wrestling with any of the problems below, get in touch directly.

    ▂▄█▄▂               ▂▄▆████▆▄▂
    prior       →        posterior
  ────────────────────────────────
  P(θ|D)  ∝  P(D|θ) · P(θ)

Bayesian Statistics

Probabilistic models that quantify uncertainty honestly — ranging from hierarchical mixed-effects models to latent variable structure. Useful when you need to pool information across groups, handle small samples, or communicate confidence ranges rather than point estimates.

Model Diagnostics · People Analytics · PyMC · GLMs · SEMs · Survival Analysis

          └——[Z]——┘
          ↓        ↓
       [X] ——————→ [Y]
     cause          effect
  ────────────────────────────
  do(X=x)  ≠  see(X=x)

Causal Inference

Identifying what actually drives outcomes, not just what correlates with them. I work with structural causal models, quasi-experimental designs, and observational data to support decisions where you can’t run a clean experiment.

DAGs · Instrumental variables · Difference-in-differences · Discrete Choice Models · Propensity scores · Experiment Design

  A/B → Δμ significant?    →  ✓
  But: network effects?       →  ?
       novelty decay?         →  ?
       long-run CATE?         →  ?
  ────────────────────────────
  test design ≠ test validity

Experimentation & Marketing Mix

Helping teams move past significance theatre toward experiments that actually inform decisions. I also build Marketing Mix Models that identify where spend is working — and where budget can be reallocated for better return.

A/B & multivariate testing · Bayesian MMM · PyMC-Marketing · Statsig · Quasi-experiments · Experiment strategy

  ┌——————————————————————┐
  │  y  ~ Normal(μ, σ)  │
  │  μ   = α + β·x      │
  │  σ  ~ HalfNormal    │
  └——————————————————————┘
     ○  ○  ○    ○  ○  ○

Training & Workshops

Hands-on sessions for data science and analytics teams — from half-day introductions to multi-day workshops with worked examples in your domain. Past topics include Bayesian workflow, causal inference, and probabilistic programming with PyMC.

Half-day team sessions from €4,500 · In-person or remote · Beginner to advanced.

Bayesian workflow · Causal inference · PyMC · Experimentation strategy

Who typically hires me

Heads of data science who need senior Bayesian or causal inference expertise without a full-time hire
Growth-stage or enterprise teams building or stress-testing their experimentation function
Analytics teams who need to upskill on causal methods or probabilistic modelling
Companies running A/B tests at scale and hitting the limits of frequentist significance testing

Get in touch

Email me at nathaniel.forde@gmail.com or reach me on LinkedIn and Bluesky. You can browse my talks and writing to get a sense of how I approach problems before reaching out.