Jibrin Jaafaru | West African Center for Emerging and Infectious Diseases

[Last updated: 16-03-2025 at 15:02 pm]

About Me

Hi! I’m Jibrin. I love building projects and following my curiosity. I am always curious about how our world operates, and whether we can find a theory that will explain it all... I am currently building RedFLow Analytics to ensure that no life is lost due to preventable blood shortages or inefficient transfusion decisions. I am using AI-powered predictive analytics to proactively bridge the gap between critical need and timely intervention to empower clinicians with data-driven insights to save lives, optimize blood resource allocation, and bring hope to patients in need.

Education

2022 - M.Sc.(With Distinction) in Artificial Intelligence and Data Analytics from the Institute of Digital Technologies at Loughborough University London
2019 - B.Sc.(Hons) in Computer Science from University of Jos, Nigeria.

Work Experience

Previously, I served as a senior Machine Learning Engineer at Greysoft Technologies, Nigeria, leading a team of three ML engineers. Together, we tackled local challenges in education, agriculture, and the food supply chain, which led to the creation of impactful startups like AgroBays Agrofoods , 5Minutes , and the BRIT App (Bookkeeping, Receipting, Invoice, and Tax).

I will be starting as a Research fellow at the West African Center for Emerging and Infectious Diseases (WAC-EID) under the D43 TRAIN grant where I will contribute to research efforts in strengthening health systems and improving infectious disease control strategies in Nigeria (Particularly Lassa Fever) My goal is build real time integrated surveillance systems that address the impact of climate change on the emergence and spread of Zoonotic Pathogens.

I have also contributed to a number of research initiatives, including projects funded by New Venture Fund through Data.org and Microsoft, the Embassy of France in Nigeria's Solidarity Fund for Innovative Projects, and the NIH-supported K43 project (K43TW011416) at JUTH.
Thanks a lot to the support of my advisors and my sponsors, I am able to conduct research and continue learning/building things that interest me.

Research Interests

I am interested in Bayesian modelling, Uncertainty Quantification, Computational Epidemiology, and Explainable Forecasting of infectious diseases. I am currently learning and exploring research papers on Simulation-based inference for Epidemiological dynamics. Feel free to recommend me some good papers in this area

Publications and Projects

EC'20

Towards increasing cervical cancer screening uptake and reducing mortality rates in Nigeria.

Jibrin Jaafaru *, Nentawe Gurumdimma, Innocent Emmanuel, Nyam Chuwang (*Main author)

The Twenty-First ACM Conference on Economics and Computation (EC'20), 2020.

PDF Conference Poster [Accepted]

21LLP008

Collaborative project with WeAreTheNews startup

As the unit team lead for the collaborative project between Loughborough university London and WeAreTheNews startup, I worked on 1. The integration of blockchain in news to revolutionise transparency, ownership, and monetization. 2. Enhancing news article credibility by ensuring the immutability of news articles. 3. Smart contracts to automate fair revenue distribution 4. Proposed the WATN coin to enable micropayments for precise content consumption. 5. Developed measures to ensure GDPR compliance.

Project Page Collaborative Project

21LLP131

Advanced Deep Q-Network for Precision Irrigation Scheduling in Northern Nigeria

I developed an advanced Deep Q-Network (DQN) model for optimal irrigation scheduling in Northern Nigeria with more complex environmental variables. The environment simulates factors such as soil type, crop type, historical weather data, and the growth stage of crops using advanced techniques like Double DQN, Dueling DQN, and Prioritized Experience Replay to improve learning stability, reduce overestimation bias, and prioritize important learning experiences.These enhancements allow the agent to more accurately predict the optimal irrigation levels needed for various crop and soil conditions, ultimately leading to more efficient water use and improved crop yields.

Code Reinforcement Learning coursework on Precision Irrigation

Gibbs sampling with a 2D Gaussian distribution: Optimizing Crop Growth with Soil Moisture and Fertilizer Levels

This article explores the use of Gibbs sampling, a Markov Chain Monte Carlo (MCMC) method, to optimize crop growth by refining estimates of soil moisture and fertilizer levels, which are crucial interrelated variables in agriculture. Gibbs sampling is introduced as a solution for dealing with complex joint probability distributions that are difficult to sample from directly, especially when variables are highly dependent

Project Page Crop Monitoring

21LLP501

Robust identification of fall army worm infestation on maize fields of small African farmholders.

This study focuses on the robust identification of fall army worm infestation in maize fields owned by small African farmholders. The research aims to develop effective methods for accurately and reliably detecting the presence of fall army worm infestations, addressing a significant threat to the agricultural productivity of small-scale farmers in Africa. The findings of this study are crucial for implementing timely and targeted interventions to mitigate the impact of fall army worm infestations on maize crops, thereby contributing to the resilience and sustainability of smallholder farming in the region.

PDF MSc Dissertation

21LLP133

Semi-supervised fabric defect detection based on image reconstruction and density estimation.

In this project I present a semi-supervised approach to fabric defect detection utilising image reconstruction and density estimation techniques. I used the existing AITEX and DAGM 2007 datasets and a convolutional autoencoder was trained to learn the normal patterns of non-defective fabrics. The autoencoder is then used to reconstruct fabric images, with reconstruction errors serving as indicators of potential defects. To enhance defect detection, Kernel Density Estimation (KDE) we applied to model the distribution of reconstruction errors, enabling the identification of anomalies based on density scores. In this way I was able to effectively detect fabric defects with limited labelled data. I evaluated the model using accuracy, precision, and recall metrics, yield 87.5, 80.2, 78.7 respectively showing promising results in identifying defective fabrics.

Code Advance Big Data Analytics coursework

MTH308

Mathematical Modelling of Anti-Malaria Drug resistance.

I focused on developing a comprehensive model for anti-malaria drug resistance. This involved a thorough examination of parasite population dynamics, drug treatment impacts, host immune responses, and various epidemiological factors. My findings contributed to the formulation of effective strategies for combating drug-resistant malaria.

PDF Mathematical Modelling coursework

Profession Membership

I am a member of ACM Special Interest Group (SIGecom) on Economics and Computation , Effective Altruism Network (EA Nigeria & EA Global). I am also affiliated with Black in AI and Data Scientists Network formerly, Data Science Nigeria.

Misc

1) For events fliers/invitation, you can download my potrait image here
2) A quote that has stayed with me all through life and work is "There is no passion to be found playing small - in settling for a life that is less than the one you are capable of living." by Nelson Mandela
3) Here are few person that I am most inspired by (in no particular order): Scott Alexander, Tyler Cowen, Paul Graham, Naval Ravikant, and Gwern.

I appreciate your interest in reading this snippet. Feel free to say hello or ask me anything using my email before you leave.