At CloudPlexo, I led the design and development of Threadline, a real-time fraud detection, anti-money laundering, and compliance intelligence platform for banks and fintechs. It combines explainable rules and machine learning to assess transaction risk while transactions are in motion.
I also contributed to AgentSpec, an AI-powered workflow automation platform, and built AI agents with the Strands Agents SDK and Amazon Bedrock AgentCore. My work includes developing client proofs of concept and delivering AI and cloud consulting from discovery through implementation and handover.
At Crunch Garage, I built a churn prediction engine for banking, insurance, telecommunications, and SACCOs, and developed an AWS cost optimization agent for analysis, anomaly detection, reporting, and forecasting.
At Reality AI, I led a squad of 20+ software engineers building AI tools and Marvel AI, a generative AI teaching assistant. I also cut page load time by 25%; earlier at Smollan, I worked on real-time fraud detection, sales forecasting, and predictive models that reduced customer churn by 15%.
