Vlad is available for hire

Vlad Constantinescu

Verified Expert in Engineering

Data Science Researcher and Developer

Corbeanca, Ilfov County, Romania

Toptal member since July 16, 2020

Expertise

Machine Learning Data Mining NLP Deep Learning C Java .NET Python Computer Vision Windows Development Microsoft Development

Bio

Vlad is a Ph.D. candidate and a senior data science researcher/engineer with eight years of experience in statistical machine learning/deep learning. He also has 15 years of experience in software engineering—ranging from agile fast-prototyping teams to large teams involved in mission-critical automotive development. Besides heaps of knowledge, expertise, and experience, Vlad also has a persistent interest in theoretical computer science, information theory, and algorithmic information theory.

Portfolio

Getvisibility

Apache Tika, Java, .NET, Apache Kafka, Docker, Data Science, Python, TensorFlow

Institute of Biochemistry of the Romanian Academy

C, Docker, Data Science, Python, TensorFlow

TotalSoft SA

JavaScript, Python, C, C++, C#, .NET, Android, Java

Experience

Computer Science - 17 years
Object-oriented Programming (OOP) - 15 years
Software Engineering - 15 years
Data Science - 10 years
Machine Learning - 8 years
Generative Pre-trained Transformers (GPT) - 5 years
Deep Learning - 5 years
Computer Vision - 5 years

Availability

Part-time

Preferred Environment

Ubuntu Linux, Debian, Microsoft Visual Studio, Windows, IntelliJ IDEA, Eclipse

The most amazing...

...system I've developed is a deep neural network representation of a graph of hundreds of thousands of protein-protein interactions and ontology relationships.

Work Experience

Machine Learning Researcher

2018 - PRESENT

Getvisibility

Developed (Python, TensorFlow, Apache Kafka) an unstructured text classification solution using a deep neural network—including second-order prediction of its own confidence, which was deployed to customers worldwide.
Built a Python-based structured parsing library using trees of regular expressions to perform complex parallelizable operations on large text corpora, which was used to feed the natural language processing pipeline.
Researched and developed (Python, TensorFlow) a method of representing documents as language-agnostic vectors using deep semi-supervised learning, which allows transfer learning between different languages for the mainline classification problem.
Developed (Python, TensorFlow) scalable automation solutions for several machine learning workflows using technologies such as Docker, Nginx, DVC, and Nomad;
Held various training seminars on theoretical and applied machine learning topics for audiences in Ireland.

Technologies: Apache Tika, Java, .NET, Apache Kafka, Docker, Data Science, Python, TensorFlow

Machine Learning Engineer

2016 - 2020

Institute of Biochemistry of the Romanian Academy

Researched and developed (C, Python, TensorFlow) a solution for embedding massive protein-protein interaction graphs (collected from sources such as BioGRID) into Euclidean geometries using noise-contrastive estimation with random walks.
Researched and developed (Python, TensorFlow) a solution for predicting secondary protein structure from primary structure with state of the art accuracy.
Researched and developed (Python, TensorFlow) a method of using a low-dimensional Euclidean embedding for smoothing the discriminative learning problem of inferring which genes are responsible for the process of cellular aging.
Researched and developed (Python, TensorFlow) a solution for controlling the trade-off between bias and variance of deep representations of massive protein-protein interaction graphs by using a novel type of variational autoencoder.
Co-authored a patent request for a novel, computationally light method of addressing the vanishing gradient problem problem in neural networks using squashing functions.
Researched the possibility of using a stochastic process over the SIR model to build a random field model which can be trained to predict the probability of COVID-19 breakouts.

Technologies: C, Docker, Data Science, Python, TensorFlow

Software Engineer

2009 - 2017

TotalSoft SA

Researched and developed (Java, Android Java, JavaScript) a prototype for a next-generation vehicle telematics platform with advanced fleet tracking capabilities, including statistical simulation workflows.
Developed (PHP) a set of libraries over own templating engine used to automatically generate C and Java 6 interfaces and wrappers from Franca IDL descriptions, including D-Bus interfaces.
Built (C, Java, Android Java) a mission-critical vehicle telematics platform with telematics services, deployed in VW vehicles worldwide. As part of this large project, I was also responsible for the core interop vehicle abstraction layer.
Developed (C#, Phalanger, PHP) a Windows Forms application using complex event processing and code generation to automatically reverse-engineer regression tests for Software Items based on their previous log output.
Constructed (ARM7 C) a cooperative multitasking OS core for the display unit of a Bluetooth car-kit (novero The Truly One), with a display and rendering stack and an API based on cooperative state machines.

Technologies: JavaScript, Python, C, C++, C#, .NET, Android, Java

Lead Developer

2004 - 2009

CallStreams Ltd.

Researched and developed with C and C# a next-generation call-recording solution over USB for digital telephony switchboards.
Developed with C and C# a suite of apps (XTR Reporter Pro) that managed analog and digital USB call recordings along with integrating various contact management apps, reporting and scoring, performing waveform analysis, and various conversions.
Developed with Borland C++ Builder a suite of interop libraries managing call recording systems over analog lines, including reverse-engineering unknown audio formats .

Technologies: C, C++, C#.NET

Experience

Getvisibility

https://www.getvisibility.com/

I applied research and handled the full-stack software development for the commercial deployment of a massively scalable natural language processing (NLP) pipeline using deep learning towards evaluating the semantic content of large volumes of unstructured data on corporate networks.

Gerontomics — System Biology of Aging

https://www.biochim.ro/group-systems-biology-of-aging/

I applied theoretical research and handled the full-stack software development for a project analyzing massive networks of genes with a deep-learning angle towards understanding which gene expression pathways are involved in the process of aging.

Education

2009 - 2011

Master of Science Degree in Artificial Intelligence

University Politehnica Bucharest - Bucharest, Romania

2002 - 2009

Engineer's Degree in Computer Science

University Politehnica Bucharest - Bucharest, Romania

Skills

Libraries/APIs

TensorFlow

Tools

IntelliJ IDEA, Microsoft Visual Studio, Apache Tika

Languages

Python, Java, C#.NET, C, C#, C++, JavaScript

Paradigms

Object-oriented Programming (OOP)

Frameworks

.NET

Platforms

Eclipse, Windows, Debian, Ubuntu Linux, Docker, Apache Kafka, Android

Other

Computer Science, Software Engineering, Machine Learning, Data Science, Data Mining, Natural Language Processing (NLP), Deep Learning, Generative Pre-trained Transformers (GPT), Mathematics, Computer Vision, Information Theory

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