Client Onboarding Process
Our 10 step process from here to full-scale machine learning capability.
(ETA: maximum of 6 months.)
Submit a brief!
It all begins with curiosity.
Are you wondering if your workflows, research or service could be improved by algorithm driven software?
Because it probably can.
Whether you know exactly what to do or simply want to understand your options, nowhere else will you be able to receive a wide variety of approaches in a single response.
Review the proposals and select the best approach, then meet your point of contact
You will be sent the best consolidated Proposals, which Suppliers have returned to your Brief. You choose the best one.
Then you meet your assigned Consultant, who have a variety of backgrounds. Whether you're a scientific or commercial practicioner, you'll be matched accordingly.
This ensures they can help you understand the suggested approach in your language!
Sign the Order Confirmation
To meet your Supplier, you will have to review our General Terms of Business and Payment Terms. You then sign an Order Confirmation, which outlines the Scope of Work.
A technical conversation
Our Brief has been designed to address project management concerns before they become concerns, asking you for all key information required for smooth development.
Often this means ensuring your technical requirements can be met by the Supplier in question. We'll arrange an hour long conversation between you and a representative from the Supplier.
This can be useful for all technical people to get on the same page and ensure there will be no hard barriers to successful deployment.
Stage 1 - Strategic Review
Most Proposal Scopes include a Strategic Review. Before committing to the build, you can chat to the recommended supplier on a technical level.
The stages can vary a little from one provider to the next, however it's common for stage 1 to involve an on-site visit (or e-visit, #pandemic) by the provider.
All potential approaches, costs and timelines, as well as the level of involvement required from your business will be discussed.
Some Suppliers offer this for free, some charge for it. It will be outlined in the Proposal Scope agreed in the previous stage.
It's well worth it.
Stage 2 - Proof of Concept
Once you've decided on your target capability, the engineers will build you a working model in a sandboxed environment using historic data.
It's a scaled down version simply to demonstrate that the final full scale model will work.
This is necessary because you can never quite know if the data a model is trained on holds the patterns needed for prediction or automation.
A machine learning solution will work only if the data holds these patterns.
Stage 3 - Full Scale Build
Most machine learning projects fail not because of the algorithms but due to the engineering.
Data Scientists are not machine learning engineers.
It's an incredible array of skills to build the requisite architecture for a working model and for these models to be deployed back into the wild correctly (into your labs or onto your website, etc.).
Our process has been designed to maximise engineering success.
When engineers improve their algorithms for one client, they often improve them for all their previous clients, too.
Why? Because they can. Community capitalism at its best.
So long as architectures on your end haven't changed, most of our engineering providers guarantee ongoing algorithm maintenance to ensure their software never breaks.
The maintenance required is entirely dependent on the complexity of the final build. If there are costs associated, it's always outlined explicitly in the Scope of Work agreed at each stage.
You can supercharge your business or research with machine learning.