Maku Mazakpe

5+ years shipping production AI systems. Specializes in multi-LLM pipelines, eval-driven development, cost-optimized serverless infrastructure, and Forward Deployed Engineering for emerging-market constraints.

Currently architecting and operating a 1,583-row AI-generated catalog pipeline (this site) on ~$61/mo of GCP infrastructure, with 71 unit tests, zero production downtime in the last 90 days, and a 41% MoM cost reduction shipped via BigQuery-driven optimization.