ML / Data Science Consultant
Summary
Short-term consulting engagement (not full-time) for a senior data scientist to review Quill's PodPromote podcast ad-strategy data, recommend the simplest approach (rules, classical ML, or LLM-assisted) hitting 85-90% explainable accuracy, define an evaluation framework, and hand off a build plan to their team. Stack: Python, scikit-learn, XGBoost/LightGBM.
At Quill, we are building PodPromote, a system that turns a podcast client's intake into a paid advertising strategy: which ad partners to use, budget per tactic, and expected downloads. We are looking for a senior data scientist or ML practitioner for a short, one-time engagement. This is not a full-time hire. We need an expert to review our data and tell us exactly what to build, including whether we need ML at all.
What you'll do
- Review our historical intake-to-strategy data and identify which inputs actually drive strategy selection
- Recommend the simplest approach (rules, classical ML, or LLM-assisted) that hits 85–90% accuracy with explainable outputs
- Define the evaluation framework and a clear build plan
- Hand it off to our team so they can own it
You have
- Production ML experience: models real clients used
- Strong small-data, tabular skills (multi-label classification, feature importance, Python, scikit-learn, XGBoost/LightGBM)
- The judgement to recommend "no ML" when that is the right call
- Clear communication and comfort mentoring a developer
- A base in Canada (required)
Details: Remote · Starting October 2026 · Hours scoped together · Rate open
Interested? Send a short note mentioning your interest in this contract position with your CV and we will get in touch with you!