About Me

Who Am I?

I am a Data Scientist with over 4 years of experience in Research, Private, and Financial Industries. My expertise encompasses a wide range of data-related skills and knowledge.

Educational Background

I hold a Master's degree in Mathematical Sciences with a specialization in Big Data and Financial Mathematics.

Certifications

I am certified by IBM and Microsoft in various data-related fields, including Data Analysis, Data Science, and Machine Learning.

Passion for Data Science

My journey into Data Science was inspired by my undergraduate studies in Mathematics, where I discovered the power of mathematical modeling to solve real-world problems. The integration of computers into the process, particularly through MATLAB and later Python, allowed me to realize my calling as a data scientist.

Skills and Expertise

• Data Skills: Proficiency in data collection, data management, data cleaning, data analysis, data visualization, machine learning, deep learning, credit scoring, and database management.
• Programming Languages: Expertise in Python, R, SPSS, SQL, MATLAB, and Julia. Exploring Java for personal projects.

Notable Project

One of my recent projects involved applying supervised machine learning to predict the comfort of individuals using autonomous cars. We employed a variety of algorithms to tackle this challenge. We used the following machine learning algorithms for prediction:

1. Logistic Regression
2. Support Vector Machines
3. K-Nearest Neighbors
4. Naive Bayes
5. Decision Trees and Random Forests
6. Neural Networks

Random Forest Algorithm emerged the best algorithm with an accuracy of 92.2%. Additionally, in collaboration with Prof. Alois Pichler (Professorhip of Financial Mathematics, Faculty of Mathematics, Technical University of Chemnitz), we were able to generate and publish a Julia package: ScenTrees.jl which is used for generating scenario trees and scenario lattices for multistage stochastic promgramming. We used stochastic approximation to provide algorithms for these processes. Additionally, we are providing a method of generating trajectories from an array of data using conditional density estimation. For more information, please visit the Package's documentation at https://kirui93.github.io/ScenTrees.jl/stable/ . The publication can also be found at https://doi.org/10.21105/joss.01912.

On my daily work, I am involved in developing credit scoring models that are currently being used in a leading financial industry to award customers credit limits. These credit models have performed exceptionally well in terms of distinguishing good and bad customers that are joining the product. We have been able to reduce the NPL rate from ~5% to ~2%, hence the business is able to make profits. I have also built different behavioral models for the business. These models have been used to increase limits to customers who are performing well in the product.