Courage Ekoh

  • Data Scientist | Machine Learning Researcher | Sustainability PhD Student @ RIT

AI for Sustainability

Hello, I'm Courage

I am a machine learning engineer/researcher interested in applying artificial intelligence to sustainable projects. Previously, I have worked in fintech as well as in the energy space and hope to continue working on projects that directly impacts the needs of the average human. I am skilled in various programming languages, building ML models from scratch and with existing libraries, MLOps and foundational software engineering

Get In Touch

Education


Aug 2022 - Present
PhD @ Rochester Institute of Technology
Sustainability & AI
Aug 2020 - Dec 2021
Master's degree @ Carnegie Mellon University
Information Technology

Concentration: Applied Machine Learning

Jun 2014 - May 2019
Bachelor's degree @ University of Ibadan
Industrial and Production Engineering

Skills


Programming

Python, Java, C/C++, SQL

Data Science

NumPy, Pandas, Matplotlib, Seaborn, Scipy, StatsModel, Tableau, Metabase, SQL, A/B Testing, MS EXCEL, Google Data Studio

Machine Learning

Sci-Kit Learn, Pytorch, Tensorflow, Tensorflow Data Validation, Hugging face, Flask

Platforms

PostgreSQL, MySQL, AWS, Git, Docker, Linux, DevOps

Work History


Aug 2022 - Present
Research Assistant @ INSYST Lab RIT, Rochester, NY

• Working with Dr. Nathan Williams on electrification problems in Sub-Saharan Africa. Current Project in Nigeria, to estimate the latent demand for electricity in the region

Jun 2022 - Aug 2022
Data Scientist (Research Intern) @ Plentify, Cape Town, South Africa

• Came up with proprietary ML algorithms for flow detection on home heaters with potential to save the company over 40% installation cost on every HotBot installed in the homes of customers

Sep 2021 - Jan 2022
Data Scientist @ Koa Technology, Nairobi, Kenya

• Worked with the operations team to provide insight from the data produced by all departments and foster growth in the company

• Maintained and updated back-end queries to constantly meet business needs in terms of the data pulled out

• Created over 90% of the data engineering processes as well as the various reporting and visualizations used by the various departments in the daily running of the company

Sep 2021 - Dec 2021
Machine Learning Consultant @ TFE Energy, Cape Town, South-Africa

• Utilized Machine learning to build a yield prediction algorithm for mini-grid connections in rural Africa

• Evaluated the possibility of data-backed refinancing for mini-grid connections by combining clustering, and classification algorithms

• Worked extensively with geospatial data to understand how development affects electricity access and consumption in rural villages.

May 2021 - Sep 2021
BI Developer & Data Analyst @ Koa Technology, Nairobi, Kenya

• Conducted statistical and data analysis with captured data and provide business intelligence to aid decision processes for the Operations, Marketing and Tech departments.

• Achieved 2,650% increase in active savers within 3-months from launch by pioneering data-driven initiatives on customer acquisition and retention.

• Built tools from ground-up that provided 100% visibility into KPIs across the entire company.

Aug 2017 - Feb 2018
Well Services Maintenance Engineer (Internship) @ Schlumberger, Port-Harcourt, Nigeria

• Deployed GPS trackers and Maintenance Information Systems on over 200 oilfield assets (covering about 60% of the project, within a month) for better asset and maintenance management as part of the Global Traceability Project

• Maintained and certified pressure control equipment such as treating lines and valves to provide up to 99.99% availability on the field

• Conducted routine maintenance on oilfield equipment, i.e., pumps, engines, mixers, etc. that ensured downtime was near 0 hours.

Research Papers


  • Kolawole, Adekunle, and Courage. O. Ekoh. "A multivariate model to predicting vibration features for equipment prognosis." Nigerian Journal of Technology 41, no. 4 (2022): 739-749.
  • Oshingbesan, Adebayo, Courage Ekoh, Germann Atakpa, and Yonah Byaruagaba. "Extreme Multi-Domain, Multi-Task Learning With Unified Text-to-Text Transfer Transformers." arXiv preprint arXiv:2209.10106 (2022).
  • Oshingbesan, Adebayo, Courage Ekoh, Chukwuemeka Okobi, Aime Munezero, and Kagame Richard. "Detection of Malicious Websites Using Machine Learning Techniques." arXiv preprint arXiv:2209.09630 (2022).

Some Projects


Fraud Detection

This project trains a fraud detection classifier for monetary transactions using a number of techniques for anomaly detection (i.e isolation forest and autoencoders) and deploys the model as an API to production using the Heroku server.

MLOps @ Coursera

This is a repo containing all assignments submited in the MLOps specialization on Coursera