AI Consultant | Artificial Intelligence & Data

• Oct 2023 - Present

Junior AI Consultant | Artificial Intelligence & Data

• Sep 2021 - Sep 2023

As a contracted ML engineer and AI developer, I optimized and implemented various ML solutions. Thus far, I’ve:• Improved the maturity of a leading Dutch e-classified ads platform’s ML systems by systematically introducing MLOps best practices. My end deliverables included: the introduction of coding and documentation standards, a MLOps maturity gap analysis, and a personalized MLOps improvement plan.• Increased system robustness of a leading Dutch e-classified ads platform’s "For You" homepage recommender. I achieved this result through extensive testing and creating a test data generation pipeline to “refresh” stale test data. All my contributions were done within the context of a very senior Data Science Team (minimum of 8+ years of development experience).• Launched a self-paced, ML-focused coding training website to standardize and improve code quality across Deloitte NL. This resource has been so effective that it is expected to be released soon to other Deloitte NSE (North and South Europe) member firms.• Designed the architecture for large-scale model storage and experiment tracking across different development environments for a German steel conglomerate. This work enabled a clean method for managing thousands of trained ML models at once.• Built a production-grade recipe recommendation system and a user order validation service for an e-commerce startup. These deliverables powered the client’s online web shop store front. Show less

Graduate Research Intern

• Jan 2020 - Aug 2020

As a research intern, I collaborated closely with the Crunchr team to integrate AI technology into their core product offerings. My primary focus was to build upon the groundwork laid by previous researchers who had identified that Crunchr’s data needed additional representation learning to be useful for machine learning. This meant encoding their highly heterogeneous relational database data in a way that most ML algorithms can "understand". My final solution: • Translated heterogeneous relational database entity information into low dimensional vector representations (embeddings), making them compatible with a wide range of machine learning algorithms. • Provided a one-off solution to data pre-processing and representation • Utilized intermediary graph structures and graph neural networks (GNNs) to ensure interpretability in the data encodings • Quantified the significance of "network" information for specific tasks Show less

Artificial Intelligence Intern

• Jun 2019 - Jul 2019

I established the technical foundation for automating cardiovascular disease risk prediction in clinical research settings. According to the literature, excessive epicardial fat is associated with an increased risk in heart failure and other cardiovascular diseases. As a result, I translated a medical researcher’s request to automate epicardial adipose tissue volume measurements into an end-to-end neural image segmentation pipeline. The designed computer vision algorithm segments cardiac fat from muscle in patient CT scans and serves as a tool for subsequent medical studies. Show less

Math and Science Tutor

• Sep 2015 - Dec 2017

I tutored 60+ students in an assortment of Applied Mathematics, Chemistry, and Electrical/Biomedical Engineering courses. I used my previous 5 years of private tutoring experience to improve students’ grades up to one letter. I tutored 60+ students in an assortment of Applied Mathematics, Chemistry, and Electrical/Biomedical Engineering courses. I used my previous 5 years of private tutoring experience to improve students’ grades up to one letter.

Computer Vision Intern

• Jun 2017 - Jul 2017

I developed a robust and computationally efficient customer detection algorithm that worked on a live video stream. To achieve this goal, I had to apply various image processing and machine-learning techniques to improve the accuracy of customer detection. I developed a robust and computationally efficient customer detection algorithm that worked on a live video stream. To achieve this goal, I had to apply various image processing and machine-learning techniques to improve the accuracy of customer detection.