print("Mauro's Portfolio")


Hello there!
I'm Mauro, a PhD student in Machine Learning at the University of Bristol (UK). I am passionate about AI and its limitless applications.

Mauro's Portfolio


Hello there!
I'm Mauro, a PhD student in Machine Learning at the University of Bristol (UK). I am passionate about AI and its limitless applications.

BIO



Hi, my name is Mauro Comi. I am a PhD student in Machine Learning at the University of Bristol (UK) supervised by Prof Nathan Lepora and Dr Laurence Aitchison. My research interests lie at the intersection of 3D Deep Learning, Neural Fields, and Robotics Perception (Computer Vision, Tactile Sensing). I have a soft spot for Computer Graphics and Physically-Based Rendering. Previously, I worked as a Machine Learning research engineer in autonomous driving at the Netherlands Organisation for Applied Scientific Research (TNO), where I developed Deep Reinforcement Learning applications for self-driving vehicles.
I obtained my Double Master's degree in Data Science at the Eindhoven University of Technology and Universidad Politecnica de Madrid. I did my Master's thesis on Deep Reinforcement Learning for importance sampling in physically-based rendering.

To know about my research, papers, and talks:

Go to my academic profile

Machine Learning and Side Projects



This is a partial list of my Machine Learning and Data Science projects completed in my free time or for academic courses. Main findings and challenges are highlighted and reported using R Notebooks or IPython Notebooks. My professional projects are not included.


Card image

TacSDF

Implementation of the TacSDF workshop paper on 3D shape reconstruction using vision-based tactile sensing.


  • PYTHON
  • PYTORCH
View More
Card image

3D Shape Reconstruction using Vision Based Tactile Perception

Reconstructing 3D object shapes using Graph Neural Networks and robotics tactile sensing.


  • PYTHON
  • PYTORCH3D
  • PYBULLET
View More
Card image

Monte Carlo Path Tracer

Implementation of different standard Monte Carlo approaches in path tracing, and comparison with Next Event Estimation


  • C++
View More
Card image

Q-Learning for Monte Carlo importance sampling

Reinforcement Learning-based approaches applied to the light transport equation for importance sampling


  • C++
View More
Card image

Deep Reinforcement Learning for Games

Deep Q-Learning agent to play the game Snake. Optimized it using Bayesian Optimization


  • PYTHON
  • PYTORCH
  • TENSORFLOW
  • DEEP REINFORCEMENT LEARNING
  • BAYESIAN OPTIMIZATION
View More
Card image

Artificial Intelligence meets Art: Neural Transfer Style

Deep Learning technique to transfer the style of a painting to a chosen image


  • PYTHON
  • TENSORFLOW
  • DEEP LEARNING
  • KERAS
View More
Card image

Monte Carlo Tree Search: Tic Tac Toe

MCTS strategy to play the game Tic tac Toe


  • PYTHON
  • MCTS
View More
Card image

Flight Delay Prediction

Modeling and predicting the delay of 7 million US domestic flights using Big Data Tools.


  • SPARK
  • SCALA
  • R
View More
Card image

Highway patrol

Java application using Flink, implementing functionalities of Speed Radar, Average Speed Control and Accident Reporter on data streams.


  • JAVA
  • FLINK
  • DATA STREAM
View More
Card image

TPC-C Benchmark

Java program using HBase to create, load tables, and implementing four different queries to simulate the activity of any company that manages, sells, and distributes products or services.


  • JAVA
  • HBASE
  • HADOOP
View More
Card image

Raw data processing for diseases prediction

Processing raw data to feed a machine learning model using IBM SPSS Modeler.


  • PYTHON
  • RAW DATA
View More

About me


  • PhD in Machine Learning (CDT in Interactive AI)

    09/2021 - present

    University of Bristol (Bristol, United Kingdom)

    A four-year PhD program in Machine Learning. I am interested in 3D Deep Learning, robotics perception, and physics simulation.

  • Master thesis researching Q-Learning and Deep Reinforcement Learning for Importance Sampling in Physically-based rendering.

    02/2019 - 10/2019

    TU Delft (Delft, The Netherlands)

    Technical University of Eindhoven (Eindhoven, The Netherlands)

    Master thesis researching Q-Learning and Deep Reinforcement Learning for Importance Sampling in Physically-based rendering.
    Supervisor: Prof. Elmar Eisemann - TU Delft
    Supervisor: Prof. Decebal Mocanu - TU Eindhoven

  • EIT Digital Double Degree - MSc in Data Science, Distinction

    09/2017 - 10/2019

    Entry University: Universidad Politecnica de Madrid (Madrid, Spain)

    Exit University: Technical University of Eindhoven (Eindhoven, The Netherlands)

    A two-year Master double degree programme, coordinated by the European Institute of Innovation and Technology (EIT). Major in Data Science, and Minor in Innovation and Entrepreneurship.
    Topics covered:
    • Machine Learning algorithms and Deep Learning (Computer Vision, NLP)
    • Big Data Ecosystem: creation of Big Data application with Spark (Scala), Flink, HBase
    • Advanced statistics with R
    • Innovation and Enterpreneurship, focused on start-ups development.

  • Premaster in Data Science and Business Information Technology

    09/2016–07/2017

    University of Twente

    Enschede, The Netherlands

    Bridge program on Computer Science and Business IT. Among the courses I took, I followed joint courses with Delft University and Standford University.

  • Bachelor of Science in Mechanical Engineering

    09/2011 – 02/2015

    Polytechnic University of Milan

    Milan, Italy

    Thesis Project: "Mechanical design and virtual prototyping for thermal turbomachinery"
    • Cost accounting and decision making, economic criteria for the dimensioning of industrial plants and facilities, layout planning
    • C++ and Matlab programming
    • Operations Research and linear planning for optimization
    • Design methods and CAD Softwares (Autodesk Inventor, SolidWorks, SolidEdge, PTC Creo)

  • Postgraduate Teaching Assistant

    02/2021 - present

    University of Bristol

    Bristol, United Kingdom

    Teaching Assistant for the units:
    • Introduction to AI, MSc unit, 2022-2023
    • Introduction to AI, BSc unit, 2022-2023
    • Introduction to AI, MSc unit, 2021-2022

  • Machine Learning Research Engineer

    03/2019 - 08/2021

    TNO - Netherlands Organisation for Applied Scientific Research

    Helmond, The Netherlands

    Focus on researching and applying Machine Learning techniques for autonomous driving:
    • Deep Learning for trajectory prediction and decision making
    • (Inverse and Deep) Reinforcement Learning for planning and decision making
    • Bayesian statistics for uncertainty modelling, Explainable AI and Interpretable AI

  • Machine Learning Intern

    03/2018 - 08/2018

    Zinkcloud

    Madrid, Spain

    Developed Deep Learning algorithms for Image detection in aerospace applications (Python, Tensorflow, Keras).

  • Internship as Product manager assistant and Aerospace Engineer

    02/2015 - 04/2015

    Jenoptik GmbH

    Hamburg, Germany

    • Mechanical design and optimization for aerospace systems (drones, Airbus components)
    • Product Management: market analysis for military and civil helicopters, market research and analysis for unmanned aerial veichle

  • Trainee as Mechanical design Engineer

    09/2014 - 01/2015 (3 Months)

    Stilmas

    Milan, Italy

    Computational Fluid Dynamics and turbomachinery optimization

About me


  • PhD in Machine Learning (CDT in Interactive AI)
    09/2021 - present

    TU Delft (Delft, The Netherlands)

    University of Bristol (United Kingdom)

    A four-year PhD program in Machine Learning. I work on 3D Deep Learning, robotics perception, and physics simulation.

    Master thesis: Deep Reinforcement Learning for importance sampling in Physically-based rendering.
    02/2019 - 10/2019

    TU Delft (Delft, The Netherlands)

    Technical University of Eindhoven (Eindhoven, The Netherlands)

    Master thesis researching Q-Learning and Deep Reinforcement Learning for Importance Sampling in Physically-based rendering.
    Supervisor: Prof. Elmar Eisemann - TU Delft
    Supervisor: Prof. Decebal Mocanu - TU Eindhoven

  • EIT Digital Double Degree - MSc in Data Science
    09/2017 - 10/2019

    Entry University: Universidad Politecnica de Madrid (Madrid, Spain)

    Exit University: Technical University of Eindhoven (Eindhoven, The Netherlands)

    A two-year Master double degree programme, coordinated by the European Institute of Innovation and Technology (EIT). Major in Data Science, and Minor in Innovation and Entrepreneurship.
    Topics covered:
    • Machine Learning algorithms and Deep Learning (Computer Vision, NLP)
    • Big Data Ecosystem: creation of Big Data application with Spark (Scala), Flink, HBase
    • Advanced statistics with R
    • Innovation and Enterpreneurship, focused on start-ups development.

  • Premaster in Data Science and Business Information Technology
    09/2016–07/2017

    University of Twente

    Enschede, The Netherlands

    Bridge program on Computer Science and Business IT. Among the courses I took, I followed joint courses with Delft University and Standford University.

  • Bachelor of Science in Mechanical Engineering
    09/2011 – 02/2015

    Polytechnic University of Milan

    Milan, Italy

    Thesis Project: "Mechanical design and virtual prototyping for thermal turbomachinery"
    • Cost accounting and decision making, economic criteria for the dimensioning of industrial plants and facilities, layout planning
    • C++ and Matlab programming
    • Operations Research and linear planning for optimization
    • Design methods and CAD Softwares (Autodesk Inventor, SolidWorks, SolidEdge, PTC Creo)

  • Postgraduate Teaching Assistant
    02/2021 - present

    University of Bristol

    Bristol, United Kingdom

    Teaching Assistant for the units:
    • Introduction to AI, MSc unit, 2022-2023
    • Introduction to AI, BSc unit, 2022-2023
    • Introduction to AI, MSc unit, 2021-2022

  • Machine Learning Research Engineer
    03/2019 - Present

    TNO - Netherlands Organisation for Applied Scientific Research

    Helmond, The Netherlands

    Focus on researching and applying Machine Learning techniques for autonomous driving:
    • Deep Learning for trajectory prediction and decision making
    • (Inverse and Deep) Reinforcement Learning for planning and decision making
    • Bayesian statistics for uncertainty modelling, Explainable AI and Interpretable AI

  • Machine Learning Intern
    03/2018 - 08/2018

    Zinkcloud

    Madrid, Spain

    Developed Deep Learning algorithms for Image detection in aerospace applications (Python, Tensorflow, Keras).

  • Internship as Aerospace Engineer and Product Management assistant
    02/2015 - 04/2015

    Jenoptik GmbH

    Hamburg, Germany

    • Mechanical design and optimization for aerospace systems (drones, Airbus components)
    • Product Management: market analysis for military and civil helicopters, market research and analysis for unmanned aerial veichle

  • Trainee as Mechanical Design Engineer
    09/2014 - 01/2015

    Stilmas

    Milan, Italy

    Computational Fluid Dynamics and turbomachinery optimization

Contact Me


maurocomi92@gmail.com
Currently in Bristol (UK)