Almost all of my work is done under strict Non-Disclosure Agreements which I take very seriously, so in most cases I can’t provide many details of projects. However, here are some brief descriptions of current and recent projects, and testimonials from clients.
I have recently (summer 2021) completed a multi-year, part time role in an aerospace startup. I worked within a small team on a number of computer vision projects, successfully using both ‘traditional’ and deep learning approaches to detect and identify potentially fast-moving targets against noisy backgrounds. One of the systems we developed on ‘edge’ hardware (Jetson NX) was part of an airborne ‘guidance system’ package which is now entering the commercial phase.
Industrial Safety Project
I am currently (as of 2021) in a ‘technical lead’ role on a project in the industrial safety sector, working with a small team using Jetson hardware and software to provide increased safety in some uniquely challenging environments. Again, this project is now entering the commercial phase, but we hope there is much future scope for further development.
I have developed the central computer vision software component in a multi-edge-device system in the Agri sector. For commercial reasons, further details are not in the public domain at this time, but the project has been very successful so far (2021), with further improvements now in progress.
I am currently (2021) providing computer vision, and wider machine learning consulting and services to a project in the UK Property sector. We have already demonstrated the ability to ‘disrupt’ some areas of this sector by using deep-learning based computer vision techniques, alongside other AI tools some of which we are developing in-house.
Other Recent Projects
As well as the above current projects, at various points over the last few years, I’ve worked on projects:
- Using computer vision to identify items being purchased in a supermarket environment.
- Classifying a certain type of medical data to provide diagnosis of a class of disease.
- Number plate recognition, once on a smartphone app, and once on Google Glass.
- Person-counting, for retail environments particularly post-Covid.
And just briefly – projects I usually turn down…
I don’t accept all work that comes my way:
- I’m not interested in trying to predict the future value of the stock market, or other financial markets. A lot of other people are already spending (and usually losing) a lot of money doing that.
- For professional indemnity insurance reasons, I don’t usually undertake projects involving in medical data, with one exception mentioned above which was an academic, not commercial, project.
A past client, Silex Microsystems AB, commented:
Tom worked for us on a project to measure critical dimensions of a product. Tom came up with several good solutions how to solve this problem. It was easy to discuss the current progress. The resulting code was easy to read and tune on-site before releasing it to production.
A client recently commented on my work for him, quite simply:
It works like a miracle.
A neural network client commented:
A great person to work with. Very good at what he does. Delivers the agreed deliverables very quickly. Updates you frequently via email on how the work is going, and what he plans to do next. Very little effort was needed to make him understand the task that needed to be done. Has great ideas on how to improve or extened the work further. Professional and high quality work and documents are provided [5/5 stars]
Another client on a neural network project commented:
Tom is very friendly and professional. I am very satisfied with the work he has done for me. [5/5 stars]
My full profile, including client feedback, is available on