Marko Smilevski Email & Phone Number

Skopje, Macedonia

Skopje, Skopski, North Macedonia

Skopje, Skopje Statistical Region, North Macedonia

Skopje, Skopski, North Macedonia

- Led a team of three Machine Learning Engineers to develop and maintain a platform for trading crypto currencies (futures) based on ML model predictions. The platform included machine learning engine to predict future trends, bot with profitable strategies, portofolio optimization algorithms and various dashboards.- Led a team of three Machine Learning Engineers to develop a Recomendation System based on implicit data, using Neural Networks for Collaborative Filtering.

Developed, trained and maintained machine learning models in the fields of:- Semantic Segmenation for architecture and construction (UNet)- Named Entity Recognition for emails (Transformers)- Time Series prediction in crypto- Recomendation Systems for Ads in a Browser (Collaborative Filtering)

Edinburgh, United Kingdom

- As part of the collaboration with The University of Edinburgh and Amazon, I was awarded a Masters' Dissertation Project that was supervised and supported by Amazon. - Proposed two novel solutions on how to compute the similarity between two products, when each product is described with text and an image.- For the proposed architectures we used Recurrent Neural Networks, Fully Connected Neural Networks, Convolutional Neural Networks, Siamese Neural Networks and the attention mechanism.- Our work is the first project that explores Deep Learning in the field of multi-modal Record Linkage.

Skopje, Macedonia

- Developed a model that predicts the cash and cash equivalents of medium and small companies in the UK.- Implemented a pipeline for pre-processing large amounts of noisy financial data.- Created an Android application for disease identification in plants using Convolutional Neural Networks.

Skopje, Macedonia

- Created a model for generating stories for a sequence of images. The model was based on a Sequence to Sequence model whilst having two encoders and a decoder. The novelty of this model is that before it generates text it first encodes parts of the image sequence and the last sentence it produced.- The model generated long narrative and evaluative language that described the visual context of the image sequence.

Skopje, Macedonia

- Worked within a team on the development of a web application for employee management.- Implemented methods that collected working times, project issues, basic info about the employees, methods that create Excel reports and methods that measured employee performance.

Master Of Science - Ms, Artificial Intelligence

Activities and Societies: Part of the EDIntelligence - The Artificial Intelligence SocietyAchieved Distinction on Master's Thesis titled.

Bachelor'S Degree, Computer Science And Engineering, 9.25/10, Cumulative Gpa: 3.7, Second-Class Honours, Upper Division (2:1)

Ranked in the top 3% of my graduate class Notable courses: Machine Learning, Data Mining, Intelligent Information Systems, Algorithms.