Tirop Meshack.
A computer engineer based in Goseong, South Korea. I freelance as a full-stack web and mobile developer, and research deep learning for clinical waveforms at Kyungdong University. Born in Kenya.
What I'm working on
I'm in my senior year of Computer Engineering at Kyungdong University. My senior thesis was on temporal prediction of heart defects from chest X-rays, a CNN trained over the CheXchoNet release that reached AUC 0.79 to 0.84. The current research picks up where the thesis pointed: waveform-level signals, specifically non-invasive intracranial pressure estimation in pediatric neurocritical care. Bidirectional LSTM on ABP and CBFV, leave-one-patient-out across 12 patients, current best MAE 2.92 mmHg. The research page has the full picture.
On the freelance side I'm building JiraniSoko, a neighbourhood marketplace for Kenyan urban communities. Listings, jobs, motors, property, all wired up to M-Pesa. Web is live; the Android app is in beta and iOS is next.
The rest of the week is reading: PhysioNet papers on noninvasive ICP estimation, the spectral and pseudo-Bayesian model-based approaches, and whatever's new in the IEEE bioinformatics track.
A few things I've made
-
2025
Non-invasive ICP Estimation
Current research. Bidirectional LSTM on ABP and CBFV waveforms, trained on PhysioNet's pediatric neurocritical-care dataset and cross-validated against CHARIS.
-
2024
JiraniSoko
Local marketplace for Kenyan neighbourhoods. Listings, jobs, motors, property; M-Pesa payments; Android app shipped, iOS in progress.
-
2024
Multition Education
Backend developer (freelance) on Multition's AI-focused course platform. Mentor-led courses across ML, DL, NLP, computer vision, and generative AI. 500+ students, 11+ courses.
-
2024
Sunrise Drilling Limited
Web developer (freelance, ongoing). Corporate site with dynamic services, secure admin backend, and observability via Grafana.
-
2023–24
Temporal Heart-Defect Prediction
Senior thesis. CNN over chest X-rays plus temporal metadata, trained on CheXchoNet under Prof. Z. I. Mannan. AUC 0.79–0.84.
-
2023
Cardiovascular Disease Prediction
Classifier trained on the Cleveland Clinic dataset with feature engineering and model tuning. 96% test accuracy.
-
2022
Earlier client work
Sites and backends for Mamafua Agency and Teddy Visuals. Auth, dynamic content, PostgreSQL.