As well as help you create an approach to apply AI & Automation in your business we can also collaborate or create AI and Automation applications for you.

We develop primarily in Python, one of the most popular scripting languages for AI & Automation use cases. We offer full portability for hosting in your cloud of choice, or within your internal systems in the case of desktop apps. We find that creating web app based functionality tends to minimise maintenance requirements and simplify dev. ops for many use cases. It also allows you to create a suite of automation tools and AI models that can be interlinked through a consistent web browser interface minimising the need for staff to learn new systems and processes.

Data Application Development Toolkit

Our Development toolkit is structured around the three buckets of activity required to deliver a data led solution:

Data Input

This includes all of the activities required for ingesting data into the application or data storage systems including:

  • Web Scraping. Typically python based approaches through Proxy services
  • API integration. Connecting systems typically using cloud functions.
  • Data migration. Helping migrate data to warehousing or other storage mechanisms compatible with onward processing.
  • Pipeline generation & ETL. Ensuring where we are utilising existing managed data sources we have a robust and well maintained pipeline process.

For a broader view of data acquisition services see data acquisition.

Data Processing

At this level we are applying predominantly Python based data processing approaches to the data to generate our outcomes.

For example:

  • Use of bespoke ML models based around common libraries including Sci Kit learn.
  • Connection to cloud ML solutions such as Google’s Vertex AI or Amazon Sagemaker
  • Connection to Pre-Trained AI services such as Clarifai or Google Translation.
  • Automated data processing for reporting and onward dashboarding such as joining multiple data sources, data cleaning and join big data sets beyond excel capabilities.

Output/ Interface

There is no single interface to data that suits every need. We prefer to agree an output that works with your business processes, this can always evolve over time.

For example:

  • Dashboarding solutions
  • Automated email delivery of output in excel form.
  • Connection to forward systems, e.g. CRM system, through API.
  • Developing of web based interfaces so you can access your data product through an interactive but bespoke interface, including streamlit.
  • Delivery to cloud data warehouses such as Bigquery or Snowflake.

POC to Production grade solutions

Our approach is based on agile and interative development. Rather than define perfection and spend six months on a build before testing, we prefer to show quick value through proof of concept or minimum value proposition based approaches. These can often deliver business value within weeks which provides a base for building and enhancing functionality over time.

However a POC that can’t be brought to production is a waste of time so we also develop with an eye to future productionalisation, either architecting for cloud deployment within any of the major cloud infrastructures or your infrastructure of choice. Fortunately the choice of Python based solutions means you are getting a solution with access to the widest pool of support infrastructure and resource available so you will neber end up locked into a proprietary system that you cannot easily maintain and upgrade.

Monitoring and Deployment

We generally recommend deployment within one of the core cloud platforms of choice as these typically offer a suite of effective monitoring and management tools which simplify and futureproof ongoing delivery, with minimal disruptions and global capabilities if required.