A unique 1 week training and consulting engagement
A Deep Dive is an opportunity to combine work on a specific machine learning problem with strategic and technical training. Over the course of 5 days, you and your team will be able to follow the progression of a data science project, learn from an expert in the field, and get valuable strategic advice on where to go next in your AI journey.
How it works
How the process works depends on whether or not you have a fixed idea for a project.
If you do, we’ll spend our time together working on that project, from inception to deployed proof-of-concept. If you don’t, we’ll spend the first day discussing your business’ needs and selecting a project to work on.
Monday - Introduction to AI
The Introduction to AI is a highlight reel of the current state of the art in machine learning. You’ll learn about the different types of learning algorithms and we’ll discuss which of these would be best suited to your chosen project. We’ll also discuss security, encryption, privacy and ethics to make sure that the solution we develop is robust and safe.
I’ll ask the key stakeholders to briefly explain how the chosen project will impact their areas of the organisation, I’ll discuss what the project will need in terms of personnel attendance and access, and we’ll work together to develop a plan for the rest of the week.
Tuesday - Data, Data, Data
A machine learning algorithm is only as good as the data it learns from. Throughout the course of this technically-oriented day, we’ll work together together to explore your data and select features for use in our project.
The day will start with an overview of feature engineering techniques and statistical testing. Then, numbers depending, we’ll split into groups to prepare a dataset for our solution to learn from.
Wednesday - A Good Model
AI projects usually work iteratively. First we select data, then we train a model, and then we tweak the model. Using our insights from the previous day, we’ll train several models and see how they work.
We’ll discuss model tuning, and more importantly, we’ll work out when to stop tuning. We’ll cover topics like imbalanced datasets, hyperparameter tuning, cross validation, and grid search.
Thursday - Understanding the Results
After two days of working at the coalface, it’s time to step back and take a look at what we’ve built. The solution, and its metrics, will be explained to senior stakeholders and there will be lots of opportunities for Q&A.
We’ll make some final adjustments based on the feedback, and we’ll get ready for the final implementation.
Friday - Implementation and Next Steps
The first half of the Friday is set aside for implementation, we’ll deploy our model (along with any data processing tasks) to the architecture of your choice.
And then we’ll spend the second half of the day discussing what worked and what could’ve gone better. We’ll discuss any ideas for future projects, or improvements to the one we’ve just worked on, and I’ll explain how you can get started on that project given what you’ve learnt over the course of the Deep Dive week.
What you’ll get
In addition to a complete and deployed machine learning project and the dedicated focus of an AI expert for a week, you’ll also get:
- A strategic plan for future projects
- Presentations covering ML algorithms, feature engineering tips, and code samples
- 3 x 1 hour follow up calls, to use as needed
How to book
If you’re interested in booking a Deep Dive, please email me.
The cost for the full week is £7,000, but there is a 50% discount for charitable organisations.