Toon Weyens
I connect people to ideas, to solutions and to each other to
help them solve their problems.
▼ My story started with the study of energy engineering in multiple
countries. This taught me how to act like an engineer.
Afterwards, I took a deep dive into the physics of nuclear fusion,
the most promising source of energy for the future. This taught me how
to think like a scientist. It also resulted in a Ph.D. degree, and a
postdoctoral Monaco Fellowship to perform research at the outstanding
ITER Organization.
To satisfy my inner mathematician, without losing sight of what
drives me, I finally redirected my career towards Data Science and
Artificial Intelligence. This taught me how to integrate, communicate
and operate like a project manager.
After spending some high quality time at DataRobot, where I enjoyed
applying my cross-domain knowledge and competencies, I joined Dataloop,
where we are building the world’s first AI Operating System.
- Values:
-
Meaningful work | meaningful relationships | radical truth and
transparency | explore and overcome weaknesses | ownership | excellence
- Abilities:
-
Curiosity | empathic and communicative | scientific and analytical |
multicultural | proactive
- Nationality:
-
Belgian
- Date of birth:
-
1987
Experience
- Nov ’22 - Current
-
Presales Solutions Architect
Dataloop
- Oct ’21 - Nov ’22
-
Customer Facing Data Scientist, Pre-Sales
DataRobot
I connect prospective enterprise Data Science customers to the
DataRobot platform so they can achieve their goals. The focus is on the
enabling technology within the framework of business relevance.
My responsibilities include
- Drive incremental revenue through creation and maintenance of
technical and business relationships with large complex sales
opportunities
- Identify client business issues and connect them to technical
capabilities of the DataRobot platform as well as partnership
opportunities through consultative approach
- Manage communication and internal cooperation of
DataRobot resources to drive team success
- Deliver unequivocal proof of the value of DataRobot
through guided PoVs directly connected to higher level business
outcomes
- Serve as trusted partner for Data Science journeys
our customers are going through
I’ve been responsible for land-and-expand deals in companies in
automotive, utilities, insurance, etc.
DataRobot delivers a unified platform for all users, all data
types, and all environments to accelerate delivery of AI to production
for every organization. It is trusted by global customers across
industries and verticals, including a third of the Fortune 50,
delivering over a trillion predictions for leading companies around the
world.
- Jan ’19 - Aug ’21
-
Industrial Data Science Consultant
MathWorks
- Use technical credibility to assist industry and research institutes
with broad adoption of MathWorks products, work in team to connect to
relevant groups in their and our organizations
- Responsibilities:
- Deep engagements: Collaborate with strategic
enterprise customers to design and subsequently coordinate creation of
modern server-based Data Science & engineering platforms with
browser-based access and predefined connectivity to data streams &
CI/CD pipelines
- Technical advisor: Guide MathWorks customers and
connect internal and external resources
- Internal expertise: Provide colleagues with
expertise and tutoring in newer areas
- MathWorks product champion: public speaking,
organization of industry or product-specific seminars, webinars and
workshops (e.g. our
4 part Series on Deep Learning with MATLAB)
- Areas of expertise: Data Science, Internet of Things, Parallel and
Cloud Computing (AWS & Kubernetes certified), High-Performance
Computing, Enterprise Integration, Mathematical and Computational
Physics
- Jan ’17 - Dec ’18
-
Postdoctoral Monaco Fellow
ITER Organization
- Produced multiple first-author publications in quality peer-reviewed
journals
- Responsible for investigation of 3-D effects on Edge-Localized Modes
(ELM) stability
- Located at multi-billion dollar ITER
project, one of world’s biggest scientific projects
- Using numerical code PB3D, developed as part of my Ph.D. project
(see below)
Education
- Jan ’12 - Dec ’16
-
Ph.D.
Universidad Carlos III de
Madrid · Eindhoven University of
Technology · ITER
Organization
- Produced multiple first-author publications in quality peer-reviewed
journals
- Designed research project to improve understanding of
high-n instabilities
- Important class of instabilities in toroidal magnetic confinement
devices for nuclear fusion
- Developed dedicated mathematical theory [Weyens
et al, 2014, P.o.P, 21, 4]
- Designed optimized numerical code, PB3D [Weyens
et al, 2017, J.c.P, 330]
- Used modern Fortran, High-Performance Parallel Computing (HPC)
techniques
- Sep ’10 - Aug ’12
-
Master of Science - Nuclear Fusion Science &
Technology
Ghent University · Universidad Carlos III de Madrid · Université de Lorraine
- European program in the Erasmus
Mundus framework.
- Fusion Engineering & Fusion Science
- Focus on culture and language
Graduated top 5%
- Sep ’08 - Aug ’10
-
Master of Science - Energy
Engineering
University of
Leuven · Technische Universität
Berlin (exchange)
- Multidisciplinary curriculum
- Strong ties to industry
- (Thermo-)Mechanical & Electrical engineering
- Focus on economical aspects of energy
- Personal experience: exchange in TU Berlin for first half of
curriculum
Graduated top 15%
- December ’18
-
Online Course - Computability, Complexity &
Algorithms
*Udacity advanced, by Georgia Institute of Technology
- Languages, countability & Turing Machines
- Complexity: P & NP
- Algorithms: dynamic programming, FFT, maximum flow
- October ’18
-
Online Course - Bayesian Methods for
Machine Learning (part of Advanced Machine
Learning)
Coursera advanced, by Higher School of Economics
Moscow
- Expectation-Maximization (EM) algorithm
- Variational Inference & Latent Dirichlet Allocation
- Markov chain Monte Carlo
- Variational Autoencoder
- Gaussian processes & Bayesian optimization
- PyMC3, GPy, GPyOpt
Graduated with honors
certificate: link
- May ’18
-
Online Specialization - Deep
Learning
Coursera intermediate, by
Andrew Ng
- Deep learning foundations by master educator Andrew Ng
- Convolutional neural networks
- Sequence models: RNN, (Bi)LSTM, GRU
- Optimizers: Stochastic, Adam, AdaMax
- Overfitting: dropout, BatchNorm
- TensorFlow & Keras
certificate: link
- May ’14
-
Summer School - 23rd
Summer School on Parallel Computing
CINECA, Casalecchio Di
Reno
- 10 day intense graduate HPC course
- Modern parallel computing systems for computation
- HPC introduction, parallel architectures, MPI & OpenMP
- Profilers & debuggers
Projects
- Jan ’13 - Current
-
PB3D
Peeling-Ballooning in
3-D
- Part of doctoral research project
- Modern Fortran code capable of efficiently analyzing
peeling-ballooning stability of toroidal magnetic confinement
devices
- Essential for stable containment nuclear fusion
- Freely available and well-documented
for experts:
- Mathematical minimization of functional, leading to generalized
eigenvalue equation
- high-n instabilities easily excited by extreme temperature
and pressure gradients from plasma (150 million °C) to the cold reactor
walls (−260.8 °C)
- General 3-D configurations, perturbed plasma edge [Weyens
et al, 2017, J.c.P, 330]
- Postdoctoral research: investigate 3-D effects
- resonant magnetic perturbations for ELM control (RMP)
- toroidal field coil (TF) ripple
website: PB3D.github.io
source: GitHub
- Feb ’18 - Mar ’18
-
Pylgrim
Elementary Shortest Path Problem with or without Resource
Constraint
- Python & C++ implementations of promising algorithms for
Elementary Shortest Path Problem (ESPP)
- From recent publications, to benchmark and learn - currently:
- problem is NP-hard, so efficient solution is interesting
mathematical and computational problem
website: GitHub
- Jan ’18 - Dec ’18
-
Kraemer
High-frequency crypto automated arbitrage trader
- Co-creator
- Collaborative Python-based project drawing from multiple areas of
expertise
- Based on computational sciences & mathematical and physical
knowledge
- Supported by state-of-the-art cryptocurrency financial modelling
& deep learning strategies
- Oct ’18
-
Facial Composits: Finding the Suspect
Capstone project for Bayesian Methods for
Machine Learning
- Helps you with getting face of suspect in crime by generating
faces
- Employed variational autoencoder in Keras and TensorFlow to generate
face images
- Combined with Gaussian Process Optimization through GpyOpt
- Interacts with you through simple binary yes/no
questions
- Optimized program to require minimal amount of interaction
Skills
- Languages
-
| English |
|
| Dutch |
|
| Spanish |
|
| Portuguese |
|
| French |
|
| German |
|
| Italian |
|
- Sales
-
- MEDDICC ·
Sales Qualification Framework used by some world’s most successful
enterprise sales teams, including DataRobot
- Insight
Selling · Michael David Harris’ story-telling approach to
selling
- Communication
-
- Audience Focused Presentation · TU/e doctoral
training course
- Write to the Top · Julia
Bindmans’ course on customer-oriented writing skills
- Computer
-
- Linux · used
as preferred operating system
- vim · used as
preferred editor
- html,
CSS · used for this curriculum vitae (with markdown
Pandoc)
- git · standard
tool for collaborative version control
- AWS · solution architect associate certified (Associate),
Azure · some experience
- K8S · certified Kubernetes Application developer
(CKAD)
- DataRobot
· detailed knowledge gained as DataRobot employee
- Domino Data
Lab · experience with enterprise MLOps platform
- ParaView, VisIt · used as 3-D
visualization tools, combined with HDF5 and XDMF
- HDF5 with XDMF · used as
performant parallel data model for HPC simulations, both for storage and
for visualization
- Apache Hadoop, Spark · in process of
educating myself on de-facto Big Data though still limited practical
experience in the form of explorative projects
- SurfSara · example of
multiple HPC Science Gateways I worked with for scientific research
- Microsoft Office & relatives ·
enterprise collaboration tools
- SalesForce · used for
Customer Relationship Management (CRM)
- Programming
-
- MATLAB · used as powerful toolset for engineers and
scientists with full workflow experience, with expertise in Parallel and
Cloud Computing, Continuous Integration & Version Control,
Deployment, WebApps, …
- Fortran · used as de-facto language for Nuclear
Fusion HPC application, such as PB3D
- Python · used for side projects, through numpy,
scipy, pandas, cython, aioprocessing, …
- R · used for ad-hoc data analysis
- C++ · used for side projects, especially useful
paired with the BOOST library
- SQL · used to get access to data for data science
projects
- Java · limited experience, used during studies
- TensorFlow, Keras ·
PyTorch · default deep learning tools
- scikit-learn · default classical machine learning
tools
- PyMC3, GPy,
GPyOpt · used for Bayesian Machine Learning
- MPI, OpenMP · used extensively for
parallelization with HPC work
- PBS Torque, SLURM · job managers
used for cluster computing
- jupyter notebooks · used for exploratory
programming
- Bash, Make, … · experience with
linux scripting and development
- yaml · used in side-projects as simple database
solutions
- LaTeX, LuaTeX · extensive
experience formatting large documents (e.g. Ph.D. dissertation)
- Other
-
Awards, Scholarships &
Grants
- 2012-2016
-
Doctoral Scholarship
Universidad Carlos III de Madrid, CINECA
PIF scholarship for four years, summer school funding
- 2017
-
Ph.D. Research Award
European Physical Society
The Plasma Physics Division of the European Physical Society (EPS)
shall grant up to four prizes annually to young scientists from the 38
European countries associated with the EPS in recognition of truly
outstanding research achievements associated with their PhD study in the
broad field of plasma physics
website: EPS
- 2019
-
Monaco/ITER postdoctoral Fellowship
Principality of Monaco
The Monaco/ITER Postdoctoral Fellowship Program allows young
researchers to participate in one of the great scientific and technical
challenges of the 21st century and to work closely with leading experts
in fusion science and technology within a unique international setting.
The principal motivation of the research fellowships is the development
of excellence in research in fusion science and technology within the
ITER framework. Brilliance and creativity, together with understanding
of the relevance of the individual research interests to the ITER
Project, are a key requirement.
website: ITER
- 2021
-
MVP Award
DataRobot
Most valuable award for Pre-Sales
website: DataRobog
Utrecht, Netherlands · +31 615 09 80 82 · weyenst@gmail.com
pdf
version · txt
version · html version ·
source