Valcon [10/2024 – Present]

Senior Principal Data Science Consultant

Valcon is a European consulting, technology and data company based in the Netherlands, Denmark, UK, Sweden, Germany, Croatia and Serbia. Their mission is to combine premium consulting with deep technology and data knowledge to add value to our clients."

• Developed an asset monitoring system for a larger Dutch infrastructure client.
• Created and taught an advanced machine learning, data science and AI course.

Omuja (10/2024 – Present)

Founder

Omuja (Zwahili for Unity) is an innovation technology startup. It aims to bring together new emerging ideas from complex systems theory, neuroscience, design, artificial intelligence, technology, social physics to explore new applications and solutions. It’s main goal is to think more carefully about how technology, machine learning and AI are integrated in our everyday lives, optimizing for human usability, value and sustainability, rather than corporate gains.

• Sebastian - Virtual Cat Companion
• Atraxia - Psychological Survival Game

Memorial Sloan Kettering [09/2020 - 10/2022]

Data Science and Tech Research Lead

As the data science and technology research lead, I led the team to explore novel machine learning models and scientific paradigms and adopt new emerging technologies in oncology.

• Designed a virtual companion for MSK's pediatric patients, an ambient A.I. supporting the child in providing security, a bonding experience, motivation, and distraction.
• Co-created Virtual Reality experiences to relieve anxiety around medical procedures.
• Built Augmented Reality to visualize medication information through medication barcodes.
• Used Agent-based models and design simulations, combining game engine Unity, agent-based modeling, and genetic algorithms to simulate the effect of layout and architecture.
• Technologies: Unity, Virtual reality (Oculus Quest), Augmented Reality (iPhone/Android), Mixed reality (HoloLens), iOS, Android, React, Neo4J, WebGL, Nvidia Jetson, Raspberry Pi, GoPro 360.
• Machine learning: Scikit-learn, TensorFlow, Keras, Google's Trainable Machine, Google MediaPipe, Apple ARKit, Apple Health Kit, NetLogo, OpenAI GPT-3, OpenCV, Unity, Blender.

Girl Scouts of the USA [03/2019 - 09/2020, N.Y., U.S.]

Lead Data Scientist

The Girl Scouts of the USA (GSUSA) have a broad and diverse set of members. The data science team at GSUSA focused on better understanding the demographics of these girls and how this shapes their path through the GSUSA experience.

• I managed a team of data analysts to generate reports consumed by the GSUSA councils..
• Created machine learning models predicting the churn of Girl Scout members based on member demographics, overall experience, and troop diversity..
• Created an NLP software suite matching PII records across different data sources.
• Technologies: Python, Snowflake, Looker, AWS, Kubeflow, SQL.
• Machine Learning: K-means clustering, logistic regression, NLP.

SparkNeuro [01/2019-01/2020, N.Y., U.S.]

Data and neuroscience consultant

SparkNeuro offered advertisers and other content creators (e.g., Netflix) the opportunity to test their content on participants whose brain activity was recorded while watching the content.

• I led a team of data and neuroscientists to develop new ML algorithms to predict cognitive, attentional, and emotional states from EEG data.
• We developed several proprietary machine learning algorithms that could decode brain activity into emotional response and attention levels over time.
• I designed neuroimaging studies (EEG, GSR, fNIRS) to benchmark algorithms.
• I built facial expression and gaze position prediction algorithms, outperforming 3rd party APIs.
• Technologies: Python, Kubeflow, AWS, SQL.
• Machine Learning: Linear/Logistics regression, time series analysis, FFT, waveform decomposition, multilayer perceptron, and computer vision.

United Nations Global Pulse [02/2016 - 12/2017, N.Y., U.S.]

Lead Data Scientist, United Nations Global Pulse

United Global Pulse functions as a small innovation hub, helping other U.N. agencies achieve the U.N.'s sustainable development goals by harnessing emerging technologies, data science, and machine learning.

• I built deep belief convolutional nets to detect settlements from satellite imagery, predict landcover type, and predict malaria prevalence (with UNHCR and UNOSAT).
• I built natural language models for tracking emergent topics in U.N. survey responses and the sentiment toward them. (In collaboration with WFP).
• I built models using cell phone data to predict the outbreak of infectious diseases (with UNICEF).
• Technologies: Python, AWS EC2, Caffe, torch, DIGITS, NLP, HTML/JavaScript, QGis
• Machine Learning: Deep Belief Neural Networks, cDBBNs, NLP, PCA

Integral Ad Science [10/2013 - 02/2016, N.Y., U.S.]

Senior Data Scientist

Integral Ad Science provides objective metrics to online advertisers about the quality and effectiveness of websites in displaying digital ads and reaching the intended target audiences. As part of the data science team, I worked on fraud detection, automated detection of objectionable web content, and building models to understand the performance of targeted ad campaigns.

• I built neural networks to detect questionable content on the web (e.g., pornography).
• I developed algorithms to predict the viewability of digital advertisements (patent awarded).
• I Co-developed tools to measure the causal impact of ads on product revenue.
• I Developed models to predict daily user activity and monitor consumer sentiment in the U.S.
• I developed models of user purchase intent from Internet usage and activity patterns.
• Technologies: Python, MapReduce, Yarn, Pig, Spark, HBase, Hive, Impala.
• Machine Learning: Deep Belief Neural Networks, logistic regression, NLP.

New York University [07/2007 - 10/2013, N.Y., U.S.]

Research scientist

At NYU, I worked on understanding how visual stimuli are encoded in the human brain using computational models, brain imaging, and machine learning.

• I studied the neural representation of visual information in the human cortex using fMRI.
• I developed machine learning algorithms to reconstruct visual stimuli from brain activity.
• I built models of visual processing in the human brain to explain experimental data.
• Technologies: MATLAB, Java, C++, OpenGL, XCode.
• Machine Learning: Logistic Regression, Support Vector Machines, PCA.

Methods

• Data science, machine learning, and predictive modeling
• Neural networks, NLP, GenAI
• Statistics, experimental design, implementation, and analysis
• Computational modeling of systems and networks
• Robotics, Virtual Reality, Augment Reality

Algorithms

• Linear and logistic regression, K-means clustering
• gaussian mixture models, k-nearest neighbors, time series analysis
• Deep belief neural networks, convolutional neural networks, auto-encoders
• PCA/ICA, support vector machines, Bayesian models, manifold learning

Programming Languages and Tools

• Python, C/C++, JavaScript, C#
• AWS, Azure, Docker, Node
• Hadoop, Pig, Spark, H2O, SQL, HBase, Hive, Impala, Zeppelin, Scikit-learn>
• Caffe, Torch, TensorFlow
• Unity, Blender, XCode, VSCode, PyCharm
• Geospatial Information Systems (ArcGIS/QGIS)

Ph.D. in Computational Neuroscience

Utrecht University, the Netherlands [2001 - 2007]

Post-baccalaureate in Geospatial Information Systems

Penn State, USA [2015 - 2016]

M.A. in Cognitive Psychology

The University of Amsterdam, the Netherlands [1995 - 2000]

US Patent 11100529

Describing a series of methods, systems, and media for generating predicted information related to advertisement viewability - Algorithms]

US Patent 11100537

Describing a series of methods, systems, and media for generating predicted information related to advertisement viewability – Infrastructure

• Brouwer GJ, Heeger DJ. Decoding and reconstructing color from responses in human visual cortex. J Neurosci. 2009 Nov 4;29(44):13992-4003.
• Brouwer GJ, Heeger DJ. Categorical clustering of the neural representation of color. J Neurosci. 2013 Sep 25;33(39):15454-65.
• Brouwer GJ, Heeger DJ. Cross-orientation suppression in human visual cortex. J Neurophysiol. 2011 Nov;106(5):2108-19.
• Brouwer GJ, van Ee R. Visual cortex allows prediction of perceptual states during ambiguous structure-from-motion. J Neurosci. 2007 Jan 31;27(5):1015-23.
• Brouwer GJ, van Ee R, Schwarzbach J. Activation in visual cortex correlates with the awareness of stereoscopic depth. J Neurosci. 2005 Nov 9;25(45):10403-13.
• Brouwer GJ, van Ee R. Endogenous influences on perceptual bistability depend on exogenous stimulus characteristics. Vision Res. 2006 Oct;46(20):3393-402.

Orange's Data 4 Development Challenge [2004/2015]

As a contributor, I built models using cellphone records, satellite imagery, and other remotely sensed data to aid Senegal's agricultural and infrastructural development.

Standby Task Force [2015 - present]

Volunteering as a digital humanitarian during several global crisis responses.

Weill Cornell Medical College [2013 - 2017]

Consultanting with Weill Cornell Medical College and leading the development of an iPad application for use in pediatric autism research-->