Consulting
I am a data science generalist. As such, any step in the traditional data science lifecycle, from gaining understanding of the business problem through model deployment to performance monitoring, is right up my alley. I work well together with domain experts, create the right visualizations to convey key information to stakeholders, and have the technical know-how to write, package and optimize codebases around the data science process.
Working on an interesting project? Let me know!
I am particular keen to help you with the following:
Data strategy
Not sure whether your available data is of sufficient quality to achieve a certain goal, or don't know what data to collect, and where to store it? I can help you build short-, mid- and long-term strategies to make your data accessible.
Data pipelining and structure
Knowledge scattered across Jupyter Notebooks accessing csv/excel files across different people's file systems? I can help design a sane and simple process that will make your knowledge more accessible, and your work reproducible.
Modelling
Using the same package and ML algorithm (e.g. scikit-learn's Random Forest) for everything, or in doubt about using accuracy as your primary evaluation metric? I have a good grasp of many modelling approaches, keep myself informed on the state-of-the-art, and can help you connect your business problem to the right metrics.
"Sanity check"
Not 100% sure about that one crucial preprocessing step, or whether using a big neural network was really necessary? I have a critical eye that may help you validate choices made anywhere between the framing of the business problem in technical terms and deployment.
Code/performance improvement
Deployment infrastructure bills too high, or not hitting prediction-time targets? I have a keen eye for inefficiencies in code (and process), and can find and fix bottlenecks throughout your codebase.