Retail Sales Forecasting
Forecasting US retail sales (FRED RSXFS series) with SARIMA, Prophet, and LSTM side by side, deployed as an interactive Streamlit dashboard.
launch app →> whoami
AI/ML Engineer @ DeepTrack.io
I build applied machine learning systems that ship — from deepfake detection models running in production, to forecasting dashboards and retrieval-augmented tools grounded in real data.
// projects
Forecasting US retail sales (FRED RSXFS series) with SARIMA, Prophet, and LSTM side by side, deployed as an interactive Streamlit dashboard.
launch app →A conversational AI agent that offers supportive, LLM-backed mental health check-ins and guidance, deployed with Streamlit.
launch app →Terrain-aware flood risk mapping — models elevation and drainage patterns to flag flood-susceptible zones.
launch app →An English/Kiswahili retrieval-augmented chatbot that answers questions grounded directly in the text of the Constitution of Kenya.
launch app →// experience & leadership
Building deepfake detection systems for a Nairobi-based company, including a Swin Transformer classification model, rPPG-based liveness detection for live calls, and cross-modal content provenance tooling.
Leading AI/ML programming for Google Developer Groups on campus — talks, workshops, and applied ML projects for students.
Running the university's Data Science and AI Club, growing membership and driving hands-on ML learning.
Managing finances and supporting technical events for the IEEE CASS student branch chapter.
// certifications
Applied Data Science Lab
WorldQuant University
AWS Educate Machine Learning Foundations — Training Badge
Amazon Web Services Training and Certification
5-Day AI Agents Intensive Course with Google
Kaggle / Google
Machine Learning Specialization
Stanford Online / DeepLearning.AI — Oct 2025
Data Science and AI Bootcamp
Lux Tech Academy — May 2025