Bridging data, science & strategy 🚀 Machine Learning 🛠️ Tool Development 📦 R Software 🧭 Leadership 🧬 Life Sciences Domain Expert ✨ Director of Data Science @ Cercle.ai
🔬 Domain expertise: Proteomics, biomarker discovery, diagnostics, life sciences, predictive modeling
📊 Technical tools: R, machine learning, statistics, Python, experimental design, reproducible research
💪 Strengths: Translating complexity, cross-functional collaboration, storytelling with data
"Making predictions is easy ... making accurate ones is much more difficult." ⎯ Me⎯
I love to solve problems.
Often the problem can be understanding a complex biological process, but it can also be as simple as fixing something that's broken (e.g. a door that jams, a bicycle, or even machine learning software). In particular, I like to apply my data science skills to better understand, or even solve, the problems we face.
Over the past 14+ years I have combined my statistical knowledge and Open-Source Software tools to solve complex problems in the Life Sciences proteomics (high dimensional) space. In so doing, I have created a comprehensive R-based machine learning analysis ecosystem that standardizes and enables biomarker discovery and predictive model development.
Sometimes the problem is inconsistency across teams or analysts ... thus I promote adherence of "tidy" data principles and am a strong proponent reproducible research and use of bioinformatics pipelines.
Other times the problem can be sharing results across the organization ... thus developing an Application Program Interface (API) infrastructure that enables anyone to access model results with ease.
With my teaching background, I find it important to mentor junior team members while simultaneously leading more senior members. This collaborative spirit is essential to building and effective team that delivers to stakeholders, fosters a sense of accomplishment, and drives revenue generation.
I am always open to discuss possible roles 🔭 and whether my skill set can solve problems in your space!
| Machine Learning 🚀 | Statistics 📊 | Open-Source 💻 | Software Tools 🔧 |
|---|---|---|---|
| Random forest | Regression problems | R | Linux🐧, MacOS 🍎 |
| Naive Bayes | Real-world data (RWD) | C++ | Git, GitHub  |
| Lasso regularization | GLMs | CI/CD | AWS |
| k-Nearest neighbour | Causal inference IPTW | LaTeX | BASH, GNU |
| PCA | Survival analysis | Python 🐍 | Docker 🐋 |
| Maximum-likelihood | Linear mixed-effects | Opencode | LLM/Agentic workflows |
- Execute organization's data science strategy, aligning analytics with business and clinical goals
- Instituted a culture of rigorous, reproducible analysis -- shifting the team from reactive one-off requests to disciplined workflows emphasizing data quality as the primary standard
- Deliver causal inference analyses on large-scale reproductive health data (IPTW, propensity scoring, covariate balancing) to inform clinical treatment protocol decisions
- Lead our key pharmaceutical partnerships, translating multi-arm observational results into clinical insights
- Collaborate with customers and C-suite to create framework for data-based decision making
- Architect and maintain a company-wide \R analytics ecosystem -- standardizing workflows from data ingestion through Quarto-driven client reporting
- Build and mentor a data science team of 3-5; own hiring, statistical analysis plans, code review, and technical growth
| Topic 🚀 | Thumbnail 📈 |
|---|---|
| False Discovery |  |
| Mixture Models |  |
| Logistic Regression |  |
| Naive Bayes |  |
| The Birthday Paradox |  |
| Mack-Wolfe Tests |  |
| Mixed Effects |  |
| Monty Hall Paradox |  |
| Decision Boundaries |  |
| Class Imbalance |  |
| Topic 🚀 | Thumbnail 📈 |
|---|---|
| Pitch Classifier |  |
- 💬 Favorite food: 🐟 🌮
- 📚 I am currently learning woodworking 🪵 ... I'm mostly good at making a lot of sawdust!
- 💬 Ask me about: bikes and
R... I'll talk your 👂 off! - 🚴 I'm an avid cyclist:
come say hi on
- I maintain several
Rsoftware libraries (📦) that implement statistical and machine learning techniques in biomarker discovery. Some of my popular published 📦 are: - These projects support analyses in the general Life Sciences (BioTech)
space to generate proteomic based insights in health spaces such as:
- cardiovascular disease
- liver disease (NASH/NAFLD)
- alcohol effects
- biological aging
- exercise status
- metabolic disease
- Favorite techniques:
- random forest
- logistic regression (ol' faithful)
- naive Bayes
- KKNN (nearest neighbor)
- survival analyses
- ensemble methods
- I am a proponent of the open-source software, conducting the majority
of my research/analysis via Linux toolkits, R, and the
RStudio/PositIDE. - I promote conforming to the adherence of so-called "tidy" data, a philosophy of data science designed to share underlying data structure, grammar, and format which facilitates the generation of reproducible analyses.
