minoo kim.
judoka and data scientist.
linkedin github résumé
Jump to Section: Work Experience Projects Judo
who am i?
My name is Minoo Kim and I'm currently a MSc student at UCL studying Data Science and Machine Learning. Previously, I attended UC Berkeley majoring in Data Science with a concentration in Business and Industrial Analytics. I love learning and creating to help the world through data. Whether it be through data-driven decision making or simple collaboration, I always believe that data/tech should be used for the mutual benefit of everyone!
In my free time, I love listening to music, playing drums, and doing judo (nationally ranked).
Programming Languages
Python ◦ SQL ◦ Java ◦ JavaScript
Technologies
Power BI ◦ NoSQL (CosmosDB) ◦ Git ◦ Pandas ◦ NumPy ◦ scikit-learn ◦ TensorFlow
Data Science & Machine Learning:
A/B Testing ◦ Causal Inference ◦ Statistical Modeling ◦ Deep Learning ◦ Model Evaluation ◦ Feature Engineering ◦ Time Series Analysis

Experience & Research

software engineer intern.
thermo fisher scientific
Collaborated with software and product teams to define key operational metrics, designing data flows to support actionable insights and metric tracking. Engineered an automated system health monitoring and alerting pipeline with a 2-minute refresh interval, resulting in a 15% reduction in detection latency. Developed and packaged a utility app (JavaScript, Angular) to streamline real-time data collection and analytics.
Thermo Fisher Scientific
data engineer intern.
the commons xr
Constructed high-volume ETL pipelines using SQL and Azure to process 2M+ records, reducing dashboard refresh times by 50% and expediting stakeholder reporting speed. Designed interactive, aggregated Power BI dashboards to track KPIs for XR class sessions, enabling cross-functional teams to self-serve answers. Collaborated with teams to translate usage patterns into core analytical requirements.
website
The Commons XR
factGrid cuneiform.
data science researcher
Oracc to Linked Open Data: tools and methods (i.e. Jupyter Notebooks) for making Linked Open Data triple statements in Wikidata and FactGrid from ORACC project data.
github
front-end web developer.
shop spero
Christian merchandise company aimed to produce intentional designs that spark hope through everyday wear and conversation. Developed website’s front page using React and Chakra UI to create image carousel on landing page.
website

Data Science Projects

supervised learning.
msc coursework
Investigated the sample complexity of various supervised learning algorithms (Perceptron, Winnow, Least Squares, and 1NN) in relation to input dimensions. Visualized performance metrics to understand algorithmic efficiency and scalability on high-dimensional datasets.
github
Sample Complexity Graph
graphical models.
msc coursework
Implemented Exact Inference, Mean Field Approximation, and Gibbs Sampling to analyze an Ising Model on a 10x10 lattice. Explored the trade-offs between exact and approximate inference methods for probabilistic modeling and unsupervised learning, demonstrating core competencies in handling high-dimensional distributions and Markov Random Fields.
github
modeling election outcomes.
causal inference project
Applied Causal Inference (IPW) to quantify candidate endorsement effects, estimating a 23.77% causal impact on vote percentage. Developed GLM and Decision Tree models to optimize electoral strategies, achieving 97% test accuracy.
flower classifier.
deep learning project
Flower image classifier using PyTorch, NumPy, and Matplotlib, achieving an accuracy of over 80% for classifying 100+ flower species. Implemented backpropagation to track loss and accuracy during training, optimizing hyperparameters to improve the model's performance. Updated metrics show >88% accuracy through targeted tuning.
github
gdp regression.
statistical analysis project
Utilized statistical analysis techniques to correlate GDP growth rates with factors such as infrastructure development and employment rates, identifying key factors that significantly influence overall GDP growth.
olfaction device.
research project
Designed a frugal and open-source prototype of an end-to-end device that can identify the odorant and replicate that odor across any length of distance using 'e-nose' technology. Leveraged VOC sensors, Bluetooth/WiFi modules, and Google Firebase to transmit smell information, vaporizing liquid-based odor with an Arduino microcontroller.
research paper presentation
Software & Game Development
avoirus.
2d dodge game
With the COVID-19 pandemic taking over the world, the importance of hygiene is apparent. What if viruses were falling from the sky? How would you protect yourself? Avoid the virus - Avoirus.
github download game
coruna.
2d platformer
Why are you not in quarantine? Run back to safety while avoiding infected enemies! Remember, you only have one shot to run back home.
github download game
block breaker.
iOS arcade game
Have fun trying to break every block with a single ball!
github download game

Judo

judo.
gallery
A gallery of all things judo.
see gallery