Georgi

Conpend’s Trade AI is a Trade Finance solution that fully digitises paper document flows and automates routine document checks. It increases efficiency and accuracy while supporting strict regulatory and compliance requirements.

For this reason, ML approaches are used to classify documents and extract fields, tables, stamps and signatures.

As a full-time engineer I acquired full ownership of the ML projects and my responsibilities include:

- Researching and implementing NLP and Computer Vision algorithms and methodologies

- Designing and building machine learning systems for production environments

- Continuously evaluating and improving models to ensure accuracy and effectiveness

- Containerizing and integrating trained models into the application

Conpend serves a global client base of over 35 leading financial institutions. I have worked closely with major banks such as Commerzbank and UniCredit.

Technologies used: Python, PyTorch, Keras, Natural Language Processing (NLP), ZenML, Docker, C#, SQL

I started part-time at the company at 3 days a week next to my Master's studies, contributing to the same document classification and information extraction pipelines

Worked on a cloud-based microservices platform to monitor IoT sensors in supply chains.

Implemented geofencing functionality, allowing clients to define zones on a map and receive alerts

Technologies used: Go, React.js, Python, Kubernetes

Contributed to the development of ABN AMRO’s online mortgage platform.

Developed backend features and integrated them with the frontend.

Implemented authentication using OpenID Connect.

Technologies used: C#, Vue.js, event sourcing, IdentityServer4, domain-driven design.

Responsible for the courses: Calculus, Web and Database Technologies, Information and Data Management.

Graded assignments and exams.