Fredrick Nyang'au | Backend Engineer | APIs, PostgreSQL, M-Pesa Integrations

About Fredrick Nyang'au

Fredrick Nyang'au is a Backend Engineer based in Nairobi, Kenya. He builds backend systems, APIs, and business automation using Python, FastAPI, PostgreSQL, and M-Pesa Daraja.

Bachelor of Science in Information Technology, Kabarak University. Available for Junior Backend Engineering roles, backend API development, and technical contracts.

Core Technical Stack

Python, FastAPI, PostgreSQL, asyncpg, Redis, Docker, Nginx, MikroTik RouterOS API, and M-Pesa Daraja payment gateway integrations.

Featured Projects

  • KukuFiti

    KukuFiti is a poultry management backend designed to help farmers track flock performance metrics such as Feed Conversion Ratio (FCR) and mortality rates. Deployed on Render, this backend is built with FastAPI and PostgreSQL, utilizing raw SQL queries via asyncpg and Common Table Expressions for analytics. Database-level CHECK constraints enforce valid flock states directly at the storage layer.

    Read KukuFiti Case Study

  • Frixel Connect: WiFi & Fiber ISP Billing Platform

    A multi-tenant backend platform connecting ISP billing, M-Pesa Daraja payments, and MikroTik RouterOS automation. Built with FastAPI and raw SQL via asyncpg, with database-level idempotency to prevent double-provisioning on Daraja webhook retries and background worker queues for decoupled provisioning. A multi-tenant backend system connecting billing, M-Pesa payment workflows, networking infrastructure, and asynchronous background processing. Built with FastAPI and raw SQL via asyncpg, with storage-layer idempotency to prevent duplicate processing on Daraja webhook retries and background queues for router provisioning. FastAPI and raw SQL via asyncpg, with storage-layer idempotency to prevent duplicate processing on Daraja webhook retries and background queues for decoupled network provisioning.

    Read Frixel Connect Case Study

  • KukuFiti

    A FastAPI backend and analytics system for centralizing flock-cycle records, mortality tracking, feed/expense data, and domain calculations. Deployed on Render, this backend is built with FastAPI and PostgreSQL, utilizing raw SQL queries via asyncpg and Common Table Expressions for analytics, with database-level CHECK constraints enforcing business rules.

    Read KukuFiti Case Study

Technical Writing

Fredrick writes about backend architecture decisions from his own production systems - how he structures FastAPI modular monoliths, handles Daraja webhook idempotency, and chooses raw SQL over ORMs. Published on Dev.to and CoderLegion.

Contact & Connect

Get in touch via email at nyangaufredrick443@gmail.com or view his work on GitHub, LinkedIn, X (formerly Twitter), and Dev.to.