Mark Tinega

I’m a Senior Machine Learning Engineer with over five years of specialised experience building production AI systems and distributed financial software platforms. I’m known for designing production-grade ML pipelines, optimising model inference latency, and architecting resilient data and backend systems for high-volume transaction environments.

At Safaricom PLC, I built real-time machine learning pipelines processing 5,000+ transactions per second using Python, PySpark, Kafka, and AWS, with low-latency inference for production systems. I constructed feature engineering frameworks processing 10TB of weekly data, and I integrated retrieval-augmented generation tools to improve feature discovery across 10,000+ assets using LLM fine-tuning and agentic workflows. I also reduced data processing time by 30% through automated orchestration, CI/CD pipelines, and workflow scheduling while maintaining model monitoring for drift detection.

In my work across fintech and open environments, I’ve delivered measurable reliability and speed gains—supporting digital finance platforms serving 30 million+ users across mobile money ecosystems, designing acceptance test plans that boosted validation efficiency by 25%, and mitigating integration errors by 20% through stronger architecture and code standards. As a freelance machine learning & AI engineer, I delivered end-to-end MLOps pipelines that reduced model deployment time by 80% using GitHub Actions, Docker, and cloud-based CI/CD with comprehensive benchmarking, while also improving client operational costs by up to 30% using optimisation algorithms. I bring a production-first mindset to LLM fine-tuning, evaluation frameworks, latency optimisation, and continuous model monitoring for reliable deployment at scale.