Radu Balaban, Developer in Baia Mare, Maramureș County, Romania
Radu is available for hire
Hire Radu

Radu Balaban

Verified Expert  in Engineering

Software Architecture Developer

Baia Mare, Maramureș County, Romania
Toptal Member Since
July 11, 2017

Along with a master's degree in machine learning from Georgia Tech, Radu is a software engineer with over 15 years of experience in the industry. He's passionate about bringing the latest advances in computer science to performant and reliable real-world applications.


KF Software Solutions, Inc. (via Toptal)
Material UI, Redux, React, JavaScript
Key Network Services, Ltd. (via Toptal)
IMSL Numerical Libraries, Java, Math.NET Numerics, VB.NET, .NET
Fintech Startup (via Toptal)
Matplotlib, Jupyter Notebook, Scikit-learn, Pandas, Python




Preferred Environment

Visual Studio Code (VS Code), IntelliJ IDEA, Linux

The most amazing...

...thing I've coded uses the Parzen density estimate to create painting-like time-lapse videos from still photographs.

Work Experience

Front-end Engineer

2019 - 2019
KF Software Solutions, Inc. (via Toptal)
  • Developed a new web single-page application (client) that allows the management of data related to forestry stock management.
  • Implemented a functionality such as data viewing, editing, and searching.
  • Wrote support for exporting reports in Excel, PDF or CSV formats.
  • Developed a new application that significantly improved the performance of the system compared to the previous implementation (which was a server-centric).
  • Added functionality to another part of the system: an offline-capable React mobile application.
Technologies: Material UI, Redux, React, JavaScript

Machine Learning Engineer

2018 - 2019
Key Network Services, Ltd. (via Toptal)
  • Developed an ML model that validates medical prescriptions issued at vet clinics. The model can identify the most likely mistakes in the prescription and also suggest missing items.
  • Created the interface for domain experts to validate and improve the model.
  • Added support for auditing the prescription issuing flow.
Technologies: IMSL Numerical Libraries, Java, Math.NET Numerics, VB.NET, .NET

Machine Learning Engineer

2018 - 2018
Fintech Startup (via Toptal)
  • Built a model that computes the optimal portfolio given certain risk tolerance.
  • Computed technical indicators such as SMA, Bollinger bands, and did feature engineering to enhance input data.
  • Ran experiments in order to evaluate the predictive value of different machine learning models related to securities prices.
Technologies: Matplotlib, Jupyter Notebook, Scikit-learn, Pandas, Python

Machine Learning Engineer

2018 - 2018
Consumer Internet Labs (via Toptal)
  • Implemented a neural network-based digit recognition system for a web application running on mobile devices.
  • Built a system that supports recognizing multiple digits drawn by touch on the mobile screen and the recognition process runs entirely on the client-side.
  • Ensured a superior accuracy by sending the system logs with failed digit classifications to a server written in Flask and then this data is then used to periodically retrain the network.
Technologies: Flask, Python, Scikit-learn, TypeScript

Front-end Engineer

2017 - 2017
Crypto Plus Certified (via Toptal)
  • Developed a web single-page application (SPA) client app for managing assets across multiple exchanges.
  • Implemented API connectivity—both towards the application server and towards external services (exchanges).
  • Implemented the payments flow.
  • Made the application mobile-friendly by using responsive components.
Technologies: WebSockets, SVG, Redux, React, TypeScript

Machine Learning Engineer

2014 - 2017
E&P Data Sense
  • Implemented a cross-platform application for monitoring streams of data from a variety of sensors in drilling or production oil and gas installations. The system analyzes the data and can send notifications in real-time.
  • Created an anomaly detection system for finding hazardous or deficient operation parameters in order to send automated alerts to supervisors.
  • Developed a sand detection feature that monitors the sand production in the well (sand can quickly deteriorate equipment). The detection process uses data from an acoustic sensor connected to the pipeline.
Technologies: Machine Learning, Data Science, Python, R

Software Architect

2011 - 2017
  • Led the development of a new product: a job search application based on psychological profile matching. It uses psychological testing results from the company’s main product to recommend the most appropriate jobs and career paths to users. The matching is done using machine learning techniques.
  • Worked as part of the company’s R&D team—focused on data visualization, and graph visualization in particular—that won several prizes at the International Symposium on Graph Drawing and Network Visualization for several years in a row.
  • Was actively involved in the product design for the main projects of the company.
  • Led the migration of the company’s main product from a Flex and Java 1.4 servlets application to HTML5/React.js and Spring MVC. The product—a large software suite for psychological testing—was targeted mainly towards enterprises and is currently in use at several large Dutch companies (such as ING).
Technologies: PostgreSQL, Oracle, React, TypeScript, Machine Learning, Spring, Java

Software Architect | Co-founder

2004 - 2011
Software Business Partners
  • Co-founded this outsourcing company.
  • Discussed the requirements with the clients, explained possible technology choices and proposed a development plan.
  • Implemented key areas of the projects, assigning tasks to the team members as well as creating training materials for the team.
  • Worked on a large number of projects, using diverse technologies, for both the desktop and the web.
  • Developed a screencasting streaming server for Flash— low-level C protocol work which required reverse engineering the Flash screen capture codec data.
  • Created a line of business desktop applications, implemented with .NET Windows Forms, WPF, and Silverlight.
  • Completed the most challenging project at the company—embedding the IBM Lotus Symphony in the browser on Windows and Mac. Symphony was a third-party pure desktop application without an embeddable ActiveX or applet.
Technologies: Microsoft SQL Server, C++, ASP.NET, VB.NET, C#

Fluid Calc

An open-source, multiplatform application that performs unit conversions and calculates formulas commonly used in well testing/drilling.

Painting Timelapse Videos from Photographs

A small-but-fun Python app that I wrote. It analyzes a photograph and uses the color and intensity distribution to generate a painting-like timelapse video from it.

Antialiasing using Deep Learning

This application adds antialiasing (edge smoothing) to rough black and white images. It uses a UNet based on ResNet18. Implemented using PyTorch and Fast.AI


CSS, C#, HTML, JavaScript, TypeScript, Java, R, Assembly, SQL, Python, C++, VB.NET


.NET, Spring, ASP.NET MVC, ASP.NET, Flask, Redux, Material UI


Win32 API, Windows Forms (WinForms), React, OpenCV, NumPy, Pandas, Apache Lucene, DirectX, jQuery, Scikit-learn, Matplotlib, IMSL Numerical Libraries, Node.js, PyTorch, Fast.ai


Object-oriented Programming (OOP), Data Science, Functional Programming


Windows, Oracle Database, Linux, Jupyter Notebook, Visual Studio Code (VS Code), Oracle, Google Cloud Platform (GCP)


Machine Learning, Software Architecture, Software Engineering, Computer Graphics, Artificial Intelligence (AI), Computer Vision, SVG, WebSockets, Math.NET Numerics, AWS Cloud Architecture, Deep Neural Networks, Convolutional Neural Networks (CNN), Deep Learning, Deep Reinforcement Learning, Reinforcement Learning


PostgreSQL, Microsoft SQL Server, Google Cloud


IntelliJ IDEA

2016 - 2019

Master's Degree in Computer Science (Machine Learning)

Georgia Institute of Technology - Atlanta, GA, USA

1998 - 2004

Bachelor's Degree with Honors in Computer Science

Technical University of Cluj-Napoca - Cluj-Napoca, Romania

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.


Share your needs

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

Choose your talent

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

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