Aleksandar Milchevski, Developer in Skopje, Macedonia
Aleksandar is available for hire
Hire Aleksandar

Aleksandar Milchevski

Verified Expert  in Engineering

Machine Learning Developer

Location
Skopje, Macedonia
Toptal Member Since
October 30, 2019

Aleksandar has more than ten years of combined research and development experience working with data, machine learning, computer vision, and signal/image processing. He enjoys working remotely and solving complex problems. He prides himself on being able to write clean, readable code.

Portfolio

Collab (Toptal Client)
Python, Amazon Web Services (AWS), Amazon Rekognition, Computer Vision...
Miracle Mill, GmbH
Amazon Web Services (AWS), Amazon Kinesis Data Firehose, AWS Glue...
Leybold (via Toptal)
Machine Learning, Audio Single Proccessing, Digital Signal Processing, Python

Experience

Availability

Part-time

Preferred Environment

Amazon Web Services (AWS), Google Cloud, Python, Linux

The most amazing...

...thing I have worked on is a DNA sequencing software.

Work Experience

Machine Learning Engineer

2020 - 2022
Collab (Toptal Client)
  • Led the research and development of an API that does video processing.
  • Extracted valuable information from videos using AWS.
  • Used Ruby on Rails for the back-end side of the app.
Technologies: Python, Amazon Web Services (AWS), Amazon Rekognition, Computer Vision, Audio Processing, Machine Learning, Data Science

Machine Learning Engineer

2019 - 2021
Miracle Mill, GmbH
  • Contributed to several projects using AWS (Glue, SageMaker, DynamoDB, and Lambda) and ETL using Apache Spark.
  • Led the choice, training, and validation of the machine learning models.
  • Placed the algorithms into production.
Technologies: Amazon Web Services (AWS), Amazon Kinesis Data Firehose, AWS Glue, Amazon SageMaker, Machine Learning, Python

Audio Signal Processing Engineer

2020 - 2020
Leybold (via Toptal)
  • Analyzed, processed, and classified sound recordings.
  • Used discrete Fourier transform (DFT) and other signal processing techniques.
  • Used logistic regression and other machine learning techniques.
Technologies: Machine Learning, Audio Single Proccessing, Digital Signal Processing, Python

Data Scientist

2016 - 2019
Nucleics
  • Contributed to improving and developing production-ready software in C/C++.
  • Researched cutting edge ideas in the field of genomics and DNA sequencing.
  • Implemented and tested several complex ideas, including the training and testing of deep convolutional neural networks.
Technologies: DNA Sequencing, Bioinformatics, Machine Learning, Keras, Python, R, C, C++

Computer Vision Engineer

2016 - 2017
Sentice Tech
  • Worked on anomaly detection in images.
  • Used OpenCV.
Technologies: Python

Signal Processing Consultant

2016 - 2016
ECGalert
  • Implemented the electrocardiogram (ECG) signal processing and denoising.
  • Implemented a pipelined discrete wavelet transform (DWT).
  • Researched the use of several FIR and IIR filters for the ECG denoting.
Technologies: Digital Signal Processing, C++

Machine Learning Research Scientist

2014 - 2016
NAGI
  • Worked on research and development of a state-of-the-art emotion recognition algorithm.
  • Took part in the Fifth International Audio/Visual Emotion Challenge and Workshop.
  • Led the development of a people tracking solution through wifi. Implemented the model using Apache Spark.
Technologies: Emotion Recognition, Machine Learning, MATLAB, Spark, C++, Python

Junior Teaching and Research Assistant

2010 - 2014
Faculty of Electrical Engineering and Information Technologies
  • Held auditory and laboratory exercises for several courses from the fields of digital signal processing.
  • Worked on power quality assessment. Used machine learning techniques to detect and classify the disturbances.
  • Took active participation in the research project "Algorithms for time-varying harmonic analysis for power quality assessment applicable on modern digital signal processors, ERA.NET PLUS project."
  • Utilized OpenCV to implement face detection using SVM.
Technologies: Digital Signal Processing, Computer Vision, Machine Learning, MATLAB, C

Junior Researcher

2009 - 2011
Dip team
  • Worked on the detection and quantification of the intensity of ringing artifacts in JPEG coded images.
  • Developed a robust multi-frame super-resolution algorithm.
  • Got introduced to the fundamental ideas of machine learning and solving inverse problems.
Technologies: Digital Signal Processing, Machine Learning, C, MATLAB

Multimodal Affective Analysis Combining Regularized Linear Regression and Boosted Regression Trees

A multimodal approach for affective analysis was developed that exploits features from video, audio, electrocardiogram (ECG), and electrodermal activity (EDA) combining two regression techniques, namely boosted regression trees and linear regression. Moreover, a novel regularization approach for the linear regression was proposed in order to exploit the temporal correlation of the affective dimensions. The final prediction is obtained using a decision level fusion of the regressors individually trained on the different groups of features. The promising results obtained on the benchmark dataset show the efficacy and effectiveness of the proposed approach.

Improved Pipelined Wavelet Implementation for Filtering Ecg Signals

https://www.sciencedirect.com/science/article/abs/pii/S0167865517302118
In this project Discrete Wavelet Transformation implementation of a bandpass filter for denoising an ECG signal was made. New improved version of a DWT bandpass filter was realized using a circular buffer. Time performance analysis and comparison of obtained solution vs. existing solutions was done.

Machine Learning Based Super-Resolution Algorithm Robust to Registration Errors

https://ieeexplore.ieee.org/abstract/document/5739234
In this work, a novel two-phase approach is proposed for robust super-resolution in the presence of registration errors and outliers. In the first phase, a machine learning method is used to create a weight matrix for every LR (low resolution) image indicating the presence of registration errors. In the second phase, super-resolution is performed using all of the LR images and the associated weight matrices, creating an image that is free of error artifacts.

Generic Face Detection and Pose Estimation Algorithm Suitable for the Face De-identification Problem

In this work, the problem of face de-identification in an image was tackled. The first step toward a solution to this problem was the design of a successful generic face detection algorithm that would detect all of the faces in the image or video, regardless of the pose. If the face detection algorithm fails to detect even one face, the effect of the de-identification algorithm could be neutralized. That is why a novel face detection algorithm was proposed for face detection and pose estimation. The algorithm uses an ensemble of three linear SVM classifiers. The first, second, and third SVM classifier estimates the face's pitch, yaw, and roll angle, and logistic regression is used to combine the results and output a final decision. Second, the face detection results and a simple space-variant de-identification algorithm are used to show the benefits of simultaneous face detection and face de-identification.
2013 - 2015

Progress towards a Ph.D. in Machine Learning

Faculty of Electrical Engineering and Information Technologies - Skopje, Macedonia

2009 - 2013

Master's Degree in Signal Processing

Faculty of Electrical Engineering and Information Technologies - Skopje, Macedonia

2005 - 2009

Bachelor's Degree in Electronics and Signal Processing

Faculty of Electrical Engineering and Information Technologies - Skopje, Macedonia

JULY 2020 - JULY 2023

AWS Certified Data Analytics - Speciality (prev. Big Data)

AWS

NOVEMBER 2019 - NOVEMBER 2022

AWS Machine Learning - Speciality

AWS

SEPTEMBER 2019 - SEPTEMBER 2021

Google Cloud Certified Professional Data Engineer

Google Cloud

APRIL 2019 - PRESENT

Data Engineering, Big data and Machine Learning on Google Cloud Platform

Coursera (GCP)

FEBRUARY 2019 - PRESENT

Machine Learning with TensorFlow on Google Cloud Platform

Coursera (GCP)

JUNE 2018 - PRESENT

Deep Learning

Coursera (deeplearning.ai)

Libraries/APIs

TensorFlow, Keras, Amazon Rekognition, OpenCV

Tools

MATLAB, AWS Glue, Amazon Kinesis Data Firehose, Amazon SageMaker

Languages

C++, C, Python, R

Paradigms

Data Science

Storage

Google Cloud

Industry Expertise

Bioinformatics

Frameworks

Spark

Platforms

Linux, Amazon Web Services (AWS)

Other

Machine Learning, Digital Signal Processing, Big Data, Deep Learning, Audio Single Proccessing, DNA Sequencing, Computer Vision, Emotion Recognition, Signal Processing, Electronics, Audio Processing

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring