Sjoerd De Haan

As a full stack developer I specialize in end to end solutions with analytics and machine learning.

Analytics and machine learning can bring enormous benefits to your company. To be successful, you have to figure out what works for you, for your employees and for your customers. If you manage to get that right, you can realize a big impact on your bottom line.

My job is to take you trough this process: from idea to first prototype, to full product. From experience I know this journey can be a lot of fun.

I have worked for small and big companies on algorithmic trading, computer vision, data processing pipelines, dashboards, fraud detection, financial forecasting, price optimization and recommender systems. Typically I deliver end to end solutions in close cooperation with the customer. This includes defining the problem, writing a time line, guiding a team of developers, creating a proof of concept, creating a backed infrastructure, writing libraries, creating dashboards, making front ends, training models, creating cloud solutions, writing documentation. All this work is done in close collaboration with the customer, usually in small iterations.

There are a couple of ideas that I like to bring

• Business and people first
People are very important in realizing business value. This includes employees and customers.

• Separate the known from the unknown
What works best for your company or your customer is not always known in advance

   On the other hand, a large set of problems in machine learning, computer science and statistics are solved and well known

• Eliminate uncertainty early on
In most cases, the human factor is the largest source of uncertainty. Will customers like new services or new features? How will employees make use of a new dashboard?
Figuring out what works may require some experiments

• Simplicity and modularity
Stick to simple solutions as much as possible
Write software that does one task well

• High quality software

   High quality software will pay back very quickly in complex environments that involve experiment

   Well designed software tends to scale once your product is going to to be a success