BossFx: building and operating a fintech education platform, solo

The problem

Forex education in Nigeria is dominated by hype and scattered across Telegram screenshots. I founded BossFx in 2024 to offer the opposite: structured, honest trading education — delivered through real software rather than a link-in-bio. That meant one person had to be the engineering team, the product manager, and the business.

Constraints shape everything

The defining engineering constraint was that I am one person with a business to run. Every architectural decision follows from it: choose boring, managed infrastructure; minimize the number of systems that can wake me up; never build what can be integrated. That constraint is also why this case study talks about trade-offs more than features.

Architecture

The platform as it runs today:

• Next.js (App Router) + TypeScript + Tailwind CSS — one framework covering marketing pages, product pages, and API routes, deployed on Vercel with zero server management.
• Supabase — authentication and PostgreSQL. Managed Postgres with row-level security beats hand-rolled auth every time when you are the only engineer on call.
• Flutterwave — payment processing suited to Nigerian rails, wired to automated digital product delivery.
• Brevo + Formspree — email automation and form handling as services, not code I maintain.
• Google Analytics + Microsoft Clarity — quantitative funnels plus session-level behavior, used to decide what to build next instead of guessing.

The hardest project: re-platforming a live business

BossFx first shipped in mid-2025 on a legacy website stack. By early 2026 it was limiting SEO, performance, and what I could build on top. The rebuild onto Next.js/TypeScript/Supabase had one hard rule: the business keeps running. Funnels, analytics, integrations, and search rankings all had to survive the migration. The rebuild shipped in early 2026 — new stack, preserved SEO, faster pages — and taught me more about production engineering than any tutorial could: migrations are risk management, not rewrites.

Trade-offs I'd defend

• Managed services over control. Supabase and Vercel cost flexibility, but they bought me the only resource that mattered: time. The wrong choice for a 50-engineer company is the right one for a one-engineer company.
• Integration over implementation. Payments, email, and forms are all vendors. My code is the glue and the product logic — the parts that differentiate.
• Analytics before features. Instrumentation went in early, so product decisions (which pages, which funnels, which products) came from behavior, not opinion.

What I'd do differently

Start with TypeScript and the component architecture from day one instead of earning the re-platform the hard way; write architecture decisions down as I made them (I do now); and treat my own GitHub as part of the product from the start — running a real platform in private looks identical to not running one at all.

What's next

The BossFx Trading Journal — trade logging with AI-assisted performance analysis (Next.js, Supabase, Claude API) — moves BossFx from education toward fintech tooling, and is being engineered in the open.