Vince is available for hire

Vince Jankovics

Verified Expert in Engineering

Machine Learning Developer

London, United Kingdom

Toptal member since October 18, 2018

Bio

Vince is an accomplished engineer specializing in machine learning and robotics. He excels in designing autonomous systems, leveraging AI to enhance perception and control. Fluent in Python and C++, Vince has a successful track record as a consultant, turning client goals into results. His passion for innovation drives him to continuously explore new technologies.

Portfolio

Dot Square Lab

Amazon Web Services (AWS), Azure, Google Cloud Platform (GCP), Kubernetes...

Cambridge Consultants Limited

TensorFlow, PyTorch, C++, Python, Machine Learning, Deep Learning...

MathWorks

C++, Simulink, MATLAB, Computer Vision, Image Recognition, Neural Networks...

Experience

Data Science - 10 years Machine Learning - 10 years Python - 10 years Machine Vision - 8 years Deep Learning - 8 years Docker - 6 years Google Cloud Platform (GCP) - 6 years Kubernetes - 5 years

Availability

Part-time

Preferred Environment

Zsh, Git, Linux, Emacs, Programming

The most amazing...

...system I've developed is a generative machine learning model that turned simple sketches into pieces of art.

Work Experience

AI Consultant

2019 - PRESENT

Dot Square Lab

Conducted proof of concepts (POCs) and feasibility studies for machine learning systems, providing valuable insights for project planning and execution.
Managed software engineering and DevOps for large-scale cluster systems, ensuring smooth and efficient support for AI research.
Played a key role in deploying and commercializing cutting-edge machine learning algorithms, bringing advanced AI solutions to the market.
Led a multidisciplinary team on a project to design an industrial IoT device and develop a machine learning system for data processing, demonstrating effective leadership and project management skills.
Hired and coached interns, fostering talent development and ensuring valuable contributions to various projects.
Worked closely with clients to understand their needs and tailor our AI solutions accordingly, leading to high client satisfaction and repeat business.
Contributed to various projects, demonstrating my versatility and adaptability in AI.
Stayed abreast of the latest developments in AI and incorporated relevant techniques into our work, ensuring that our solutions remained cutting-edge and effective.
Worked on various machine learning projects, from image recognition systems to predictive models, demonstrating my broad expertise in the field.
Balanced multiple projects and priorities successfully, demonstrating my ability to work effectively under pressure and deliver high-quality results.

Technologies: Amazon Web Services (AWS), Azure, Google Cloud Platform (GCP), Kubernetes, PyTorch, Python, C++, C, Machine Learning, Data Science, Deep Learning, OpenCV, Robotics, Artificial Intelligence, Quantitative Modeling, Data Analysis, Mathematical Modeling, Image Recognition, Computer Vision, MySQL, Web Scraping, Reinforcement Learning, Object Detection, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Forecasting, Neural Networks, Data Engineering, Machine Learning Operations (MLOps), Team Leadership, Convolutional Neural Networks (CNNs), Data Visualization, Pandas, Matplotlib, Ray, Docker, Ludwig, Google Cloud, Hugging Face, XGBoost, Jupyter, Financial Modeling, Programming, User Interface (UI), Integration, Generative Adversarial Networks (GANs), Machine Vision, Flask, SQL, Redis, Full-stack, Internet of Things (IoT), Research, Large Language Models (LLMs), OpenAI GPT-4 API, Data Scraping, Software Architecture, Fine-tuning, Deep Reinforcement Learning, Computer Vision Algorithms, AI Consulting, Llama 3

Machine Learning Engineer

2017 - 2019

Cambridge Consultants Limited

Improved and tailored state-of-the-art algorithms, leading to the development of advanced machine learning systems that provided cutting-edge solutions to clients.
Developed robust deep learning models using PyTorch and TensorFlow, enhancing our machine learning solutions' predictive accuracy and efficiency.
Implemented data collection and preprocessing strategies for imaging problems, improving the quality of input data for our machine learning models.
Designed and developed models for time series and sensor data classification using boosted trees and neural networks, enhancing the performance of our predictive systems.
Deployed machine learning systems on edge devices using highly optimized software developed in C++, improving the speed and efficiency of our solutions.
Developed efficient data mining and processing pipelines using Python, enhancing the speed and accuracy of our data analysis processes.
Created tools for monitoring and managing machine learning training pipelines and large datasets, improving the efficiency and effectiveness of our machine learning operations.
Interviewed and coached interns, contributing to the development of talent within the organization and ensuring successful contributions to both internal and external projects.
Worked closely with clients to understand their needs and tailor our machine learning solutions accordingly, leading to high client satisfaction and repeat business.
Stayed abreast of the latest developments in machine learning and incorporated relevant techniques into our work, ensuring that our solutions remained cutting-edge and effective.

Technologies: TensorFlow, PyTorch, C++, Python, Machine Learning, Deep Learning, Artificial Intelligence, Data Analysis, Data Science, Image Recognition, Computer Vision, MySQL, Web Scraping, Reinforcement Learning, Object Detection, Forecasting, Neural Networks, Data Engineering, Machine Learning Operations (MLOps), Team Leadership, Convolutional Neural Networks (CNNs), Data Visualization, Pandas, Matplotlib, Ray, Docker, Google Cloud, XGBoost, Kubeflow, Jupyter, Financial Modeling, Programming, User Interface (UI), Integration, Generative Adversarial Networks (GANs), Machine Vision, Flask, SQL, Redis, Full-stack, Internet of Things (IoT), Data Scraping, Large Language Models (LLMs), Software Architecture, Fine-tuning, Deep Reinforcement Learning, Computer Vision Algorithms

Application Support Engineer

2016 - 2017

MathWorks

Resolved many technical issues for customers in diverse fields such as robotics, control systems, signal processing, embedded systems, and machine learning, improving customer satisfaction and retention.
Played a key role in developing a drone simulation project, where my proficiency in C++, Matlab, and Simulink led to a successful and efficient simulation model.
Worked on unit testing new features for the deep learning toolbox, ensuring their functionality and reliability.
Applied my knowledge of various technical fields to provide effective customer support and solutions.
Contributed to a project that involved a complex multi-agent system with robotic balls (Spheros) rolling around and racing in a physical arena.

Technologies: C++, Simulink, MATLAB, Computer Vision, Image Recognition, Neural Networks, Machine Learning, Data Visualization, Programming, Integration, Machine Vision, Internet of Things (IoT), Software Architecture, Computer Vision Algorithms

Robotics Intern

2016 - 2016

DroneX

Developed software using C++ and ROS to control a UAV platform, enhancing operational efficiency.
Designed and developed the control system and simulation for a bipedal robotic system, improving its stability and performance.
Constructed a mechanical test rig for a bipedal robotic system, enabling comprehensive testing and calibration.
Worked on embedded software design, creating a real-time power management application for a low-cost microcontroller.
Applied my knowledge of robotics in a practical setting, gaining valuable experience in software development, control system design, and embedded systems.

Technologies: Embedded C, C++, Python, Robotics, Mathematical Modeling, Engineering, Programming, Integration, Machine Vision, Software Architecture, Computer Vision Algorithms, Robot Operating System (ROS)

Student Research Assistant

2013 - 2014

University of Southern Denmark

Contributed to the Embodied Motion Intelligence for Cognitive, Autonomous Robots (EMICAB) research project aimed at improving the cognitive abilities of autonomous robots.
Designed a tactile-sensitive robotic fingertip, enhancing the robot's interaction with its environment.
Developed an automated test framework in C++, streamlining the testing process and improving efficiency.
Created a machine learning model for sensor characterization, optimizing sensor performance and reliability.
Integrated my work into the larger project successfully, demonstrating my ability to collaborate effectively in a research setting.

Technologies: Python, C++, Robotics, Computer Vision, Engineering, Programming, Integration, Machine Vision, Research

Experience

AI Chatbot on Custom Data

https://dotsquarelab.com/case-studies/streamlining-internal-document-queries-with-chatgpt-and-pinecone

I worked on a proof of concept to customize a GPT-based AI chatbot using internal company data. The chatbot can process both natural language and tabular data and relies on OpenAI APIs and LangChain. The system can deliver high-quality, up-to-date answers on proprietary data.

Generative AI Platform

https://dotsquarelab.com/case-studies/scalable-deployment-of-fine-tuning-text-to-image-ai-models

I developed a platform that can be used to fine-tune generative models (e.g., stable diffusion) to custom subjects and the platform utilized Hugging Face for the modeling framework and Ray Serve for the serving and training, where the whole platform (front-end, back-end, and model pipelines) was deployed in a Kubernetes cluster; the Ray deployment leveraged Kuberay to manage the lifecycle of the different components.

Process Optimization System

https://dotsquarelab.com/case-studies/using-ai-to-generate-and-optimise-chemical-plant-design

I developed a system to optimize distillation sequences and parameters for non-ideal mixtures. The system significantly reduced the time required for process engineers to design systems that meet requirements.

Pixel-perfect Instance Segmentation

https://dotsquarelab.com/case-studies/ai-based-vehicle-cut-outs-for-online-marketplaces

I developed a deep learning pipeline for an application that required pixel-perfect segmentation masks for retail. It involved carefully selecting and combining state-of-the-art models to achieve the required performance.

AI for Drug Discovery

I contributed to an industry disruptive startup that aims to solve drug discovery with AI. I worked on software engineering and DevOps for large-scale cluster systems to support the research team and designed and implemented a system architecture that made experimentation more cost-efficient and faster. Also, I proposed architectural changes to the research team regarding generative models.

Financial Timeseries Prediction

I worked on a proof of concept for a financial forecasting model that involved large amounts of unevenly spaced time series data. I developed a PyTorch data loader interface to an SQL data warehouse to efficiently train models at scale and tailored deep learning models.

AI Artist Demo

https://www.cambridgeconsultants.com/turning-our-sketches-into-art-with-machine-learning/

I worked on a machine learning technology demonstration that turned sketches into art pieces. I built on top of cutting-edge deep learning architectures using generative adversarial networks to create a system that can hallucinate artistic paintings based on the input edges. I used Python with TensorFlow to build the architecture, implement the data pipeline and facilitate distributed training. Since the work we've done for clients was confidential, this demo could perfectly showcase the team's capability, and this demo, in particular, had a significant impact on our presence in the AI consulting domain.

Beyond Human Vision Demo

https://www.cambridgeconsultants.com/artificial-intelligence-moves-beyond-human-vision/

This technology demonstration aimed to reconstruct distorted images, so the model did pixel-wise mapping and had to fill in missing information based on the context. Generative adversarial networks are excellent at doing this, and it was shown how the model could change the output depending on the environment. I used Python with PyTorch for the training and a Flask back end with a TypeScript front end for the deployed system.

Artificial Neural Network Based Adaptive Non-linear Control

I worked on a non-linear controller that can learn the system dynamics that it controls. An artificial neural network is trained online to predict the system states to provide better performance than a simple PID algorithm. It was applied to a simple two-degrees-of-freedom robotic arm to prove the concept. I developed the prototype's mechanics and electronics, simulated the dynamics in C++, and deployed the system to a low-cost Atmel microcontroller using Embedded C.

Humanoid Walking Robot

I developed a system design for a humanoid walking robot and simulated and prototyped the control algorithm using Simulink. I implemented the Linux middleware for the actuators and sensors on the onboard controller.

IMAV 2017 Virtual Challenge

http://www.imavs.org/2017/virtual-challenge.1.html

I worked on the simulation environment for the IMAV drone competition. I used Gazebo, ROS, and Simulink to provide a high-fidelity simulation that can be used to evaluate autonomous flight controllers.

Reinforcement Learning-based Crypto Trading Bot

I developed an RL agent for cryptocurrency trading that uses OHLCV data and technical indicators, and the model has been developed using PyTorch and Ray, which enables massive scaling to reduce the training time significantly. The model was trained and backtested with data from the last few years, and it showed promising results, outperforming baseline strategies.

To find the best-performing agent, I evaluated different RL algorithms (PPO, IMPALA) and model architectures (fully connected, convolutional, attention-based).

For backtesting and live trading, I used Freqtrade.

Manufacturing Monitoring IoT Device

I developed an IoT device that can be attached to injection molding machines to monitor different characteristics based on the measured temperature and piston velocity. This data was then used to create a digital twin of the machines, which provided predictive maintenance capabilities and general insight into the machine's health.

BMS IoT Scanning and Management

I worked on an IoT scanner that specifically targets devices used in building management services (BMS) to control systems like HVAC, lighting, etc. The tool scanned the network and gathered useful information about the broad range of devices used within the building.

Custom Image Generation Pipelines

I developed custom image generation pipelines for various use cases, and the deployed solution was implemented on GCP with a custom-built serving back end based on Ray Serve. This involved working on workflows that combine the most recent methods and fine-tuning open source approaches, such as Stable Diffusion, on clients' data. The results were consistently great, exceeding clients' expectations and outperforming proprietary solutions, such as DALL-E and Imagen.

This solution ensured that the deployment scaled up when needed and also scaled down to avoid using resources when there were no users.

Education

2015 - 2016

Master's Degree in Robotics

University of Bristol - Bristol, England

2012 - 2015

Bachelor's Degree in Mechatronics

University of Southern Denmark - Sonderborg, Denmark

Skills

Libraries/APIs

Matplotlib, Pandas, OpenCV, PyTorch, NumPy, Scikit-learn, TensorFlow, XGBoost, React

Tools

Jupyter, Zsh, Oh My Zsh, TensorBoard, Plotly, Git, Spacemacs, MATLAB, Emacs, Docker Compose

Languages

C++, Python, C, Bash, SQL, Embedded C, Simulink, R, Embedded C++

Frameworks

Swagger, Ray, Flask, Boost, Spark

Platforms

Linux, Docker, Google Cloud Platform (GCP), Kubernetes, Kubeflow, Android, Amazon Web Services (AWS), Azure, Bluetooth Low Energy (LE)

Storage

Google Cloud, MySQL, MongoDB, Redis

Paradigms

Agile

Other

Data Analysis, Data Visualization, Deep Learning, Robotics, Machine Vision, Machine Learning, Data Science, Artificial Intelligence, Generative Adversarial Networks (GANs), Reinforcement Learning, Image Recognition, Computer Vision, Web Scraping, Object Detection, Neural Networks, Machine Learning Operations (MLOps), Convolutional Neural Networks (CNNs), Engineering, Generative Pre-trained Transformer 3 (GPT-3), Hugging Face, Programming, Integration, Full-stack, OpenAI GPT-4 API, Data Scraping, Software Architecture, Fine-tuning, Deep Reinforcement Learning, Computer Vision Algorithms, Architecture, Technical Leadership, AI Consulting, Generative Artificial Intelligence (GenAI), Retrieval-augmented Generation (RAG), Prompt Engineering, Quantitative Modeling, Mathematical Modeling, Robot Operating System (ROS), Condor, OOP Designs, Natural Language Processing (NLP), Forecasting, Data Engineering, Team Leadership, Generative Pre-trained Transformers (GPT), Financial Modeling, User Interface (UI), Internet of Things (IoT), Research, Large Language Models (LLMs), Minimum Viable Product (MVP), Llama 3, Unmanned Aerial Vehicles (UAV), Image Processing, Optimization, Image Generation, Simulations, FastAPI, Chatbots, LangChain, Language Models, Ludwig

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