Benjamin Efenudu

Senior full-stack engineer. I build AI products and the platforms they run on.

I owned the dashboards the Central Bank of Nigeria, 18+ commercial banks and 450+ microfinance institutions use to run a regulated blockchain payment network. Now I'm building SchoolWithEase, an AI-native school management platform. Money moving between banks, AI in the classroom: I build software that can't afford to be wrong.

Email me WhatsApp meCV GitHub LinkedIn

Benjamin Efenudu, in glasses and a black T-shirt, black-and-white portrait

What I do

Senior engineering across AI, fintech and education products: I take a product from a blank repo to production, and keep it running.

AI products people can trust

LLM features that ship to real users: retrieval-grounded assistants, agentic tool calling scoped to each user's permissions, AI drafting with human approval, and guardrails that keep personal data away from the model.

Claude API, OpenAI API, Voyage embeddings, RAG, tool calling, prompt design, LLM evaluation

Fintech at central-bank scale

Four years owning the dashboards of a regulated blockchain payment network: 400+ API endpoints for onboarding, settlement, disputes and real-time transaction tracking, on systems held to 99.9% availability.

React, TypeScript, Redux, React Query, Vite, real-time dashboards, Vitest, Cypress

Full stack, front end to infrastructure

Web apps, installable PWAs and APIs, from the component library to the database policy to the deploy pipeline. Multi-tenant SaaS on Postgres row-level security, real-time UIs over WebSockets, Web3 backends on Solana and Ethereum.

Next.js, Node.js, NestJS, PostgreSQL, Prisma, MongoDB, Docker, Kubernetes, AWS, Cloudflare, CI/CD, Playwright, Cypress

Ownership and leadership

Sole frontend engineer at Zone for nearly two years, building 80%+ of every dashboard version, and named one of the company's MVPs. Before that, I led delivery teams across web, backend and mobile, and taught frontend and backend to new and intermediate developers.

Technical leadership, roadmapping, code review, mentoring, Agile delivery, technical writing

Selected work

Products I owned end to end. The payment work is under NDA, so those pages use rebuilds on made-up data and my own diagrams.

The SchoolWithEase homepage: 'One app for the whole school', beside a register of every role in a school

SchoolWithEase2025 — Present

SchoolWithEase, an AI-native school management platform

Multi-tenant SaaS that runs a whole Nigerian school, built end to end: an LLM drafts lesson notes for teachers to approve, and RAG assistants answer only from the school's own records.

Read the case study

A rebuild of the Zone Client Dashboard home with synthetic data: a sidebar, wallet balance, eight KPI tiles and a recent-transactions table

Zone Payment Network2022 — 2026

Zone Admin & Client Dashboards

The React and TypeScript dashboards the Central Bank of Nigeria, 18+ commercial banks and 450+ microfinance institutions use to run a regulated blockchain payment network in real time.

Read the case study

Three ZonePay screens with demo data: a payment review paying a merchant from a linked bank account, the home screen with linked accounts and recent transactions, and the payment-complete receipt

Zone Payment Network2026

ZonePay PWA

One installable Next.js PWA for both sides of a payment: customers pay merchants from linked bank accounts, and merchants watch each payment land in real time over WebSockets.

Read the case study

Stylised NFT pawn platform under attack: a draining balance, an alert and a shield

Solana Pawn Stars2022

Solana Pawn Stars: stopping a live drain

Frontend lead on a Solana NFT pawn platform delivered in two weeks, and the night a live security breach was draining funds: what it took to stop it.

Read the case study

AI-native, by default

Agents do the typing. I do the thinking: the requirements, the design system, the decisions, and the workflow the agents run in.

Requirements before code: Every feature starts as written requirements, research and a plan, down to the edge cases, roles and approval rules, before an agent writes a line.
A design system to build against: Design tokens, component rules and sign-off pages give agents one source of truth, so what they build matches the design instead of inventing its own.
The decisions stay mine: Agents propose; I choose. Architecture and design decisions are written down with their reasons, so each trade-off is made once, on purpose.
Agents execute the plan: Claude Code, Codex and Cursor write code, tests and docs against the plan, and I review every change before it ships, including on a regulated blockchain payment network.
A workflow I keep refining: Rules files, a task board and a handoff log after every session, so the next agent starts where the last one stopped and lessons from one session become rules for the next.
AI in the product: An LLM drafts full lesson notes in about thirty seconds, assistants answer from a school's own records through tool calls, and a chatbot answers students from their course materials.