Ivan Vasilev, Developer in Sofia, Bulgaria
Ivan is available for hire
Hire Ivan

Ivan Vasilev

Verified Expert  in Engineering

Neural Networks Developer

Location
Sofia, Bulgaria
Toptal Member Since
November 15, 2012

Ivan is an enthusiastic software engineer and machine learning researcher. His experience ranges across many fields and technologies, but his primary focuses are deep learning, Python, algorithmic trading, and Java. Ivan has written three books and has several open-source projects.

Portfolio

SAP
Python, PyTorch, Transformers, Large Language Models (LLMs)...
https://tradestorm.io/
Python, Algorithmic Trading, Stock Trading, Event-driven Programming...
Packt Publishing
TensorFlow, Keras, PyTorch, Python, Artificial Intelligence (AI), Deep Learning...

Experience

Availability

Part-time

Preferred Environment

Git, Linux, GitHub

The most amazing...

...software I've created was one of the 1st deep learning libraries, written in Java and with GPU support.

Work Experience

Development Expert, Artificial Intelligence

2024 - PRESENT
SAP
  • Developed Large Language Model-based applications and integrations.
Technologies: Python, PyTorch, Transformers, Large Language Models (LLMs), Artificial Intelligence (AI), Machine Learning, Natural Language Processing (NLP)

Co-founder and Lead Engineer

2019 - PRESENT
https://tradestorm.io/
  • Developed a full-scale algorithmic trading platform. The platform is written in Python and deployed on Amazon AWS. It supports both live trading and backtesting.
  • Implemented, deployed, and monitored 10+ automated trading strategies in the live trading environment.
  • Implemented and deployed multiple integrations with external data providers and APIs, including Interactive Brokers, FIX protocol, Binance, IQFeed, Quandl, Alpaca, and others.
Technologies: Python, Algorithmic Trading, Stock Trading, Event-driven Programming, PostgreSQL, Pandas, FIX Protocol, Interactive Brokers API, Binance API, Amazon Web Services (AWS), Numba, Crypto, NumPy, REST, Redis, Data Science, GitHub, Trading, Data Analysis

Author

2018 - 2023
Packt Publishing
  • Wrote the thoroughly revised 2nd and 3rd editions of the book "Python Deep Learning."
  • Wrote the book "Advanced Deep Learning with Python."
  • Created open-source repositories for the three books.
Technologies: TensorFlow, Keras, PyTorch, Python, Artificial Intelligence (AI), Deep Learning, Machine Learning, Deep Neural Networks, Neural Networks, Convolutional Neural Networks (CNN), Data Science, GitHub, Language Models, Transformers, Computer Vision, Natural Language Processing (NLP)

Machine Learning Engineer and Researcher

2017 - 2019
Self-employed
  • Contributed to an open-source event-based Python algorithmic trading library. The library aims to provide an easy way of testing financial ML algorithms.
  • Built a real-time and historical bar and tick data from IQFeed.
  • Integrated an API with Quandl and INTRINIO, and other data providers.
  • Developed storage and retrieval of historical data and other datasets with PostgreSQL and InfluxDB.
  • Back-tested historical data using different datasets.
  • Implemented order placement via the Interactive Brokers Python API.
Technologies: InfluxDB, TensorFlow, Pandas, Python, Artificial Intelligence (AI), Deep Learning, Machine Learning, Deep Neural Networks, PostgreSQL, Neural Networks, Convolutional Neural Networks (CNN), Object-oriented Programming (OOP), GitHub

Machine Learning Engineer and Researcher

2014 - 2016
ExB Group
  • Participated in and won several machine learning competitions.
  • Developed an accomplished machine learning library for neural networks based on Java and OpenCL.
  • Embedded new machine learning libraries, including Caffe and Keras.
Technologies: OpenCL, TensorFlow, Keras, Caffe, Python, Artificial Intelligence (AI), Deep Learning, Machine Learning, Deep Neural Networks, Java, Neural Networks, Convolutional Neural Networks (CNN), Object-oriented Programming (OOP), Aparapi, GPGPU, GitHub

Machine Learning Engineer

2013 - 2014
Self-employed
  • Authored the first open-source Java deep learning library with GPU support as a way to introduce myself to deep learning and produce something meaningful at the same time. It is implemented with Java 8.
  • Developed the library that became a successful open-source project.
  • Continued its development in-house after it was acquired by a company (ExB).
Technologies: OpenCL, Java, Artificial Intelligence (AI), Deep Learning, Machine Learning, Deep Neural Networks, Neural Networks, Convolutional Neural Networks (CNN), Object-oriented Programming (OOP)

Mobile and Web Developer

2013 - 2013
Fanattac (via Toptal)
  • Developed a mobile version of the site using Ember.js, PHP, and REST.
  • Worked on a redesign of the desktop version of the site using Backbone.js, PHP, MySQL, and REST.
  • Integrated successfully into an existing fully-remote team.
Technologies: MySQL, Backbone.js, Ember.js, JavaScript, PHP, HTML5, SQL, Object-oriented Programming (OOP)

Founder

2010 - 2013
IGI Soft, Ltd.
  • Developed an advanced SaaS web platform for trading automotive parts online, located at Zakolite.bg/. The platform allows companies to create webshops and participate in a common marketplace.
  • Integrated the system with the most widely adopted and comprehensive auto parts database and some of the leading CRM software providers in Bulgaria.
  • Led the design, development, deployment, and support of the platform.
  • Marketed the product and worked with customers on collecting feedback and improving the site.
Technologies: Jetty, Apache Tomcat, CSS, HTML, Dojo Toolkit, jQuery, Apache Maven, MySQL, Hibernate, Apache Wicket, Java, HTML5, SQL, Object-oriented Programming (OOP)

Software Developer

2010 - 2012
Bulgarian Academy of Sciences
  • Contributed to a semantic web scientific project as part of my master's thesis: Semantic Technologies for Web Services and Technology Enhanced Learning, or SINUS, located at http://sinus.iinf.bas.bg/index.php.
  • Developed a SPARQL graphical designer and semantic annotator for the project.
  • Defended my master thesis based on the project successfully.
Technologies: Jetty, Apache Tomcat, CSS, HTML, Dojo Toolkit, Apache Wicket, OWL API, Java, OWL, RDFs, RDF, SPARQL, Artificial Intelligence (AI), HTML5, JavaScript, SQL, Object-oriented Programming (OOP)

Software Engineer

2007 - 2011
Micro Focus International
  • Developed and maintained a service-based solution with a service and service-consuming Visual Studio extension.
  • Built an online survey platform based on Google GWT.
  • Developed and maintained the application portfolio management enterprise view. This included development in C++ and the legacy code maintenance of the web application.
Technologies: Windows Communication Foundation (WCF), Windows Presentation Foundation (WPF), Visual Studio, C#, Apache Maven, Apache Struts, Google Web Toolkit, Java, SQL, Object-oriented Programming (OOP), Subversion (SVN), HTML

Junior Developer

2006 - 2007
IGE + XAO
  • Developed an electrical module for the PLM product CATIA V5 using C++ and CATIA (computer-aided three-dimensional interactive application).
  • Acted as a junior developer (this was my first official job besides internships), I successfully integrated into the existing team.
Technologies: CATIA, C++

Intern

2004 - 2005
Tara Soft, Ltd.
  • Developed websites using the LAMP (Linux, Apache, PHP, MySQL) architecture.
  • Converted a static HTML design (from a designer) into a dynamic site using PHP and MySQL, all while considering the client’s requirements.
Technologies: CSS, HTML, JavaScript, MySQL, PHP

Python Deep Learning, 3rd Edition

https://www.amazon.com/dp/B0BRKYYPPC/
I wrote the thoroughly revised 3rd edition of the book "Python Deep Learning," published by Packt. The book covers the theoretical underpinnings of neural networks and the most significant network architectures, including convolutional, recurrent networks, transformers, and large language models (LLM). The book discusses problems in computer vision and natural language processing.

Advanced Deep Learning with Python

https://www.packtpub.com/data/advanced-deep-learning-with-python
I wrote the book "Advanced Deep Learning with Python," published by Packt. The book covers the mathematical theory behind neural networks and all major network architectures, including convolutional and recurrent networks, transformers, and self-driving vehicles.

Python Deep Learning, 2nd Edition

https://www.packtpub.com/big-data-and-business-intelligence/python-deep-learning-second-edition
I wrote the thoroughly revised 2nd edition of the book "Python Deep Learning," published by Packt. The book covers the theoretical underpinnings of neural networks and the most significant network architectures, including convolutional and recurrent networks and reinforcement learning. The book discussed problems in computer vision and natural language processing.

Python Algorithmic Trading

https://github.com/ivan-vasilev/atpy
I'm the author of an event-based Python algorithmic trading library focusing on machine learning. Some of the features are:

• Real-time and historical bar and tick data from IQFeed
• API integration with Quandl and INTRINIO
• Storing and retrieving historical data and other datasets with PostgreSQL and InfluxDB
• Backtesting historical data using different datasets
• Placing orders via the Interactive Brokers Python API

I have successfully transformed this open-source project into a startup (Tradestorm.io).

First Place at Skin Lesion Analysis Towards Melanoma Detection Challenge at ISBI 2016

While working at ExB, my team and I participated in and won a medical image segmentation contest, "Skin Lesion Analysis towards Melanoma Detection challenge," at ISBI 2016. Our solution was based on a combination of convolutional networks implemented in Keras.

Second Place at Gland Segmentation Challenge Contest at MICCAI 2015

While working at ExB, my team and I participated and won second place in a medical image segmentation contest, "Gland Segmentation Challenge Contest" at MICCAI 2015. Our solution was based on a combination of convolutional networks, implemented in Keras.

Deep Neural Networks

https://github.com/ivan-vasilev/neuralnetworks
I wrote the first Java deep learning library with OpenCL support. The library included the most popular neural network architectures, including convolutional and recurrent networks. The library was acquired by ExB, where I continued its development in-house.

ZaKolite.bg

A SaaS platform that allows automotive parts companies to create webshops and participate in a common marketplace, located at Zakolite.bg. The system is integrated with the most widely adopted and comprehensive auto parts database.

Semantic Technologies for Web Services and Technology Enhanced Learning

Interdisciplinary research project (Bulgarian Academy of Sciences) aimed at advancing the two fastest evolving information technologies—service-oriented computing and technology-enhanced learning by applying the semantic web service methodology.
2009 - 2012

Master's Degree in Artificial Intelligence

Sofia University St. Kl. Ohridski - Sofia, Bulgaria

2005 - 2009

Bachelor's Degree in Computer Systems and Technologies

Technical University of Sofia - Sofia, Bulgaria

Libraries/APIs

Keras, PyTorch, Aparapi, TensorFlow, Pandas, jQuery, Songkick API, RequireJS, Rdio API, OWL API, SoundCloud API, Backbone.js, Facebook API, Dojo Toolkit, Interactive Brokers API, Binance API, NumPy

Tools

Apache Maven, Subversion (SVN), Git, GitHub, Apache Tomcat, Jetty, Visual Studio, CATIA, PyCharm, Gradle

Frameworks

Caffe, Apache Wicket, Ember.js, Hibernate, Apache Struts, Windows Presentation Foundation (WPF), OpenCL, Google Web Toolkit

Languages

Python, SQL, Java, HTML5, JavaScript, CSS3, C#, HTML, CSS, PHP, RDF, C++, SPARQL, OWL

Paradigms

Object-oriented Programming (OOP), Agile Software Development, GPGPU, Event-driven Programming, Data Science, REST

Storage

PostgreSQL, MySQL, InfluxDB, Redis

Platforms

Eclipse, Linux, Windows, Amazon Web Services (AWS)

Other

Artificial Intelligence (AI), Algorithmic Trading, Deep Learning, Machine Learning, Neural Networks, Deep Neural Networks, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNNs), Full-stack, Automated Trading Software, Data Analysis, FIX Protocol, Stock Trading, APIs, Back-end, Trading, RDFs, Windows Communication Foundation (WCF), Numba, Crypto, Transformers, OpenCL/GPU, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Language Models, Computer Vision, Large Language Models (LLMs)

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