Engineering Scalable Machine Learning Systems from Complex Data.
My career progression reflects scientific exploration across multiple data domains — seismic
and electrical resistivity surveys, oceanographic measurements, ground-penetrating radar imaging, power-consumption
time series, and multi-terabyte satellite archives. Across each domain, I transform raw data into structured representations,
develop predictive models, and implement scalable computational systems. The methodology adapts; the mathematical foundation
remains constant.
Dual Ph.D. in Computational Mathematics, Science, and Engineering (CMSE) and Earth & Environmental Sciences (EES), Michigan State University.
Quick links
Core skills
This site documents the systems I’ve built across diverse data domains, including ground-penetrating radar, satellite intelligence, and energy forecasting. Each project demonstrates how mathematical structure, computational efficiency, and scalable implementation turns complex data into predictive systems.
What I build
End-to-end predictive systems across complex data domains:
Gaussian Process regression models, spectral signal decomposition (FFT/SVD),
autoregressive forecasting systems, high-resolution landcover
classification algorithms, and distributed satellite processing
frameworks scaling to 100+ cores.
Data → representation → model → validation → deployment.
What I study
Structure discovery in high-dimensional systems:
spatiotemporal forecasting, harmonic decomposition of signals,
and scalable feature engineering.
Whether the input is radar reflections, oceanographic series, energy demand curves,
or satellite imagery — I model structures that generalize.
What I’m seeking
Data Science, ML Engineering, and AI Research roles where I can apply my data-driven modeling skills.
I build models that move beyond prototypes — into reliable, deployable intelligence.
Contact
Let’s talk
Email: johnsalako3@gmail.com
LinkedIn: linkedin.com/in/john-salako
GitHub: github.com/TheGospeler