Bespoke Web Platforms
I design and build scalable online platforms (websites, APIs, widgets, etc.), managing and maintaining all aspects of the technical infrastructure.
I design and build scalable online platforms (websites, APIs, widgets, etc.), managing and maintaining all aspects of the technical infrastructure.
Find local fitness classes near you, with a pay-as-you-go model, in addition to premium services.
The MagnaPass website was originally built by 2 developers, with one half in Ruby, the other half in Angular. Due to a lack of technical guidance, product development was too slow and over complicated.
After re-writing the entire platform in a more scalable language, I then began work on integrating credit card payments, and a partner dashboard.
Searching for classes is powered by an internal (JSON) API. This was the first logical step in preparation to offer a comprehensive API for a mobile app and other third-party developers.
During development, a framework began to emerge with API end-points naturally falling into place, this led the way organically to a mobile app and offering the service to third-party developers.
Providing price-comparison and digital advertising products to publishers.
I developed a platform to download and process over 250 million products, from 10 different partners, covering the US, Canada, Australia, and Europe. This data was then offered via a real-time API, or alternatively Javascript widgets to website publishers.
Each partner provided their data in a different format (XML, JSON, or CSV) with different structures and fields.
The platform munged this data together into a consolidated/internal structure. Processing was performed across 10 servers, storing the data in a collection of MongoDB instances.
The data was then made available in an API spread across a fully-redundent cluster of ~70 servers.
The bespoke API dealt with authentication, authorisation, rate-limiting, and logging.
The API also was geo-load-balanced between Los Angeles in the US, and Stockholm in Sweden.
Online job board, serving both candidates and recruiters.
Historically the CV-Library platform stored CV content and job listings in a MySQL database. Trying to perform full-text searches across 4 million CVs was taking too long, in some cases literally 5 minutes for results to be returned.
I built the CV-Library API to provide full-text searching across the full dataset of CVs, with response times under 50ms.
The API was then extended to support searching for jobs, and applying for them.
Finally, the API was consumed by the CV-library website, encapsulating all of the job boards' online functionality.
A "brochure" website is a small-scale site consisting of static content, such as "About Us", "Our Work", "Contact Us", etc.
