Lev Yastrebov

I created the algorithm for 3D sea bottom modeling based on the data from nautical charts. Charts are a unique source of navigational data because it has the best sea coverage. However, building a triangulated surface using raw chart data will be almost useless because charts have simplified, only essential bathymetry data. Thus, the surface will have large, flat areas. So, I read a lot of scientific articles about surface interpolating methods, tested them, and picked the most suitable ones.

Another big problem was physical conflicts in the chart data. For example, one chart point can have three levels. To overcome this, I developed a special method that resolves the conflicts and doesn't make conflicts with chart data.

Moreover, I added the self-testing mode to check and adjust the algorithm. My testing tool picked charts with different levels of detail in the same area, ran my algorithm on less detail, and then compared the results with a more detailed chart.

As a result, my algorithm made it possible to connect charts with different isobath levels to create a seamless sea model, which is used for automatic routing for ships. Also, it is helpful in oceanography, hydrology, fishing, and other applications.

I developed an example of a fully functional web API with users, roles, authentication, integration with another API, and a lot of tests. I also set up a CI pipeline, which builds and runs all the tests, including the tests against the real database.

Using this application, users can create an account, log in, and store some food records to track sugar consumption. If the user does not provide the sugar amount, the app will fetch it from a nutrition data provider. Also, users can access API to check how much sugar they ate during the current day.

I made this educational project to show how to build custom lightweight authentication and authorization without using full-blown technologies like ASP.NET Core Identity and how to integrate different things into one working and robust application.

The interval map is a data structure that associates intervals of keys with values, where the key can be any comparable type with a defined lowest limit, and the value can be anything checkable for equality.

I developed an extremely efficient algorithm for assigning values; it uses only two relatively expensive O(log N) operations to find range bounds. It utilizes precise iterator arithmetic to locate all other positions.

In that case, it's very important to carefully test this algorithm to be sure that it works properly. To accomplish that, I used test-driven development alongside property-based testing instead of example-based testing. That means checking that the object under tests has specific properties after the performed operation instead of comparing result data with examples. This technique made it easy to write clear and concise tests to cover all possible data variations.
Finally, I successfully built a complex, efficient algorithm that works correctly.