Guillem Duran Ballester, Developer in Palma de Mallorca, Spain
Guillem is available for hire
Hire Guillem

Guillem Duran Ballester

Verified Expert  in Engineering

Machine Learning Engineer and Developer

Palma de Mallorca, Spain

Toptal member since August 8, 2022

Bio

Guillem is a machine learning engineer and AI researcher with a strong passion for education and open source. He works with companies to develop innovative machine learning solutions that provide business value. Guillem enjoys helping clients understand how they can benefit from the latest discoveries in AI.

Portfolio

Self-employed
Pandas, Python, Docker, Machine Learning, DataViz, Data Science...
Miguel Hernández University
Networking, Operations Research, Research, Complex Networks, Python
Good AI
Research, Artificial General Intelligence (AGI), Deep Learning...

Experience

Availability

Part-time

Preferred Environment

Ubuntu, PyCharm, GitHub

The most amazing...

...thing I've developed is an AlphaZero-based DNN architecture search solution that achieved state-of-the-art performance on computer vision tasks.

Work Experience

AI and ML Consultant

2019 - PRESENT
Self-employed
  • Designed and implemented end-to-end machine learning solutions to automate core business processes in the marketing and sports betting industries.
  • Led and mentored a data science team and advised several companies on how to architect and develop cost-efficient AI solutions.
  • Designed and implemented novel optimization algorithms for NP problems.
Technologies: Pandas, Python, Docker, Machine Learning, DataViz, Data Science, Machine Learning Operations (MLOps), PyTorch, AWS, Scikit-learn, Generative Pre-trained Transformers (GPT), NLP, Optimization, Artificial Intelligence

Research Team Member

2016 - PRESENT
Miguel Hernández University
  • Joined the research group called Aplicaciones de los Sistemas Dinámicos Discretos y Continuos, MTM2016-74921-P (AEI/FEDER, UE), working on applications of discrete and continuous dynamical systems.
  • Helped coordinate an interdisciplinary team of university professors specializing in complex systems analysis and networking engineering.
  • Proposed new research topics and built prototypes that led to four peer-reviewed publications.
Technologies: Networking, Operations Research, Research, Complex Networks, Python

AI Researcher and Engineer

2020 - 2021
Good AI
  • Designed, implemented, and deployed a PyTorch Lightning multi-GPU training pipeline on AWS for state-of-the-art research projects, reducing training costs by 80% and allowing us to train deep learning models five times faster.
  • Built the visualization and model analysis tools that enabled our researchers to iterate faster and understand the results better.
  • Proposed improvements and optimizations to different research projects and reproduced recently published research to benchmark our findings.
Technologies: Research, Artificial General Intelligence (AGI), Deep Learning, Deep Reinforcement Learning, PyTorch, Python, Git, Machine Learning Operations (MLOps), Data Visualization, Artificial Intelligence

Senior Engineer | External Consultant

2020 - 2020
McKinsey & Company
  • Acted as the product owner of an asset management tool for the energy industry.
  • Coordinated with the non-technical stakeholders to understand their needs and designed a product to address them.
  • Implemented a working asset management prototype successfully.
Technologies: Asset Management, Optimization, Python, DataViz, Data Science, Pandas, Gurobi

Production Machine Learning Engineer

2019 - 2019
source{d}
  • Designed and implemented a tool that processed commits' history in a PySpark cluster and allowed project managers to understand how developers collaborate by providing interactive reports in an Apache Superset dashboard.
  • Maintained the machine learning stack of the company, Docker containers, and the continuous integration pipeline of the ML projects.
  • Designed and built a deep reinforcement learning framework in Keras.
Technologies: Python, Spark, DataViz, Pandas, Docker, GitHub, Apache Superset, Deep Learning, Machine Learning Operations (MLOps), Deep Reinforcement Learning, Keras, Artificial Intelligence

Research Scientist and Engineer

2018 - 2019
InstaDeep
  • Implemented different reinforcement learning algorithms for both continuous and discrete problems.
  • Designed and implemented a novel architecture search solution in PyTorch based on AlphaZero, which achieved state-of-the-art performance on computer vision tasks.
  • Mentored the PhD students, taught them about coding and documentation quality standards, and contributed to the core business algorithms.
Technologies: Python, PyTorch, Deep Learning, Deep Reinforcement Learning, DataViz, Git, Artificial Intelligence

Freelance Data Scientist

2013 - 2018
Self-employed
  • Developed business intelligence and machine learning solutions for several small companies.
  • Designed and implemented successful algorithmic trading strategies for a stock portfolio that achieved 15% more returns than the reference market index.
  • Designed and implemented the MVP of a boat rental marketplace.
Technologies: Algorithms, AWS, Data Science, Data Visualization, Pandas, Python, Machine Learning, Risk Models, Options Trading, Applied Physics, E-commerce marketing, Business Intelligence Development

AI Researcher

2016 - 2016
HCSoft Programación
  • Developed the theoretical foundations of a novel AI theory base on non-equilibrium thermodynamics.
  • Built the benchmarking and visualization tools to evaluate the performance of different optimization algorithms.
  • Developed the working prototypes to validate research hypotheses.
Technologies: Python, Optimization, DataViz, GitHub, Data Science, Machine Learning, NumPy, Pandas, Artificial Intelligence

Plangym | Library for Adapting RL Environments to Planning Tasks

https://plangym.readthedocs.io/en/latest/
Plangym is an open-source Python library for developing and comparing planning algorithms. It provides a standard API to communicate between algorithms and environments and a standard set of environments compliant with that API.

Library for Configuring MLOps Best Practices in Open-source Projects

https://mloq.readthedocs.io/en/latest/
ML Ops Quickstart is a tool for initializing machine learning projects following MLOps best practices.

Setting up new repositories is a time-consuming task that involves creating different files and configuring tools such as Lint, Docker containers, and continuous integration pipelines. The goal of mloq is to simplify that process and enable a user to start writing code as fast as possible.
2012 - 2016

Bachelor's Degree in Network Engineering

Polytechnic University of Catalonia - Barcelona, Spain

Libraries/APIs

PyTorch, Pandas, Scikit-learn, Keras, NumPy

Tools

GitHub, PyCharm, DataViz, Git, Gurobi

Languages

Python

Frameworks

Spark

Paradigms

Business Intelligence Development

Platforms

Ubuntu, Docker, AWS

Other

Machine Learning, Data Visualization, Deep Learning, Artificial Intelligence, Deep Reinforcement Learning, NLP, Machine Learning Operations (MLOps), Algorithms, Asset Management, Portfolio Optimization, Quantitative Risk Analysis, Signal Processing, Networking, Electrical Engineering, System Security, Data Science, Research, Artificial General Intelligence (AGI), Optimization, Apache Superset, Risk Models, Options Trading, Applied Physics, E-commerce marketing, Reinforcement Learning, Operations Research, Complex Networks, Generative Pre-trained Transformers (GPT)

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