Killer app to automate compiling of documents in shipping sector

An Odessa start-up created a killer application for the automatic processing of admistrative documents regularly used in shipping sector.

CODESKA LLC is a start-up founded by Andrew Lander. The startupper boasts 10 years of professional experience in maritime logistics, ship handling and chartering, port forwarding and agency services.

When I first saw industry imperfection, I tried to fix it and I was fully satisfied with the effect. That was the start of something great, and still, we have a plenty of work to do!

Andrew Lander
Andrew Lander (left), Founder of CODESKA

To create this new project Andrew created CODESKA together with a technical partner, who is today his Lead Architect, with 5 year’s experience in multimodal logistics: railway, air, sea and land transportation, warehouse logistics and forwarding.

Automated recognition of templates

The first product created by CODESKA is a an app useful for the shipping sector to automate the data entry work of administrative documents. The application adopts machine learning technologies and special neural network.

The problem

Most logistic companies, especially those working in the container transports, have bad practice of manual input of document data into their ERP system. This practice leads to human errors, which sometimes cause financial losses, and extra work, which lead to higher operational costs.

The solution

BLOR is a special software based on machine learning algorithm for automated recognition of templates. For example, BLOR works with bill of ladings issued by container lines and fully kills manual works by managers as they do not need to input data into their ERP system by hand.

A special “export” function makes it a one-click procedure with a total speed up to 15 seconds per document and highest quality recognition by fields. BLOR supports all container lines documents in pdf and jpg files and the CODESKA team is getting ready to process with home BLs, invoices and other types of templates.

We wanted to create software that not only recognizes and returns text massive, but also identifies the kind of documents and feels fields with the necessary information fast and accurately. We also wanted to make data extraction to ERP systems as easy as possible.

Andrew Lander

Of course, all these processes are secured and the service company does not store any confidential commercial information.

We decided to use machine learning technologies and special neural network for this application. Using algorithms made by our science team, we managed to get 99% of accuracy and up to 15 seconds of recognition time for the document.

Andrew Lander

Therefore, more accuracy, less risk of human mistakes and big saving of working time. This product is a good example of a merge of two important corporate cultures in Odessa: the traditional maritime sector and the fast-growing IT technology.

For more information: