Skip to main content
From Student to Freelance Software Engineer: Lessons in Building Real Software

From Student to Freelance Software Engineer: Lessons in Building Real Software

December 14, 2024 3 min read Career, Freelance, Software Engineering, Mentorship, Kenya

From Student to Freelance Software Engineer: Lessons in Building Real Software is a Developer Journal article by Ancel Ajanga on https://ancel.co.ke. A retrospective on transitioning from computer science theory to delivering production software for clients: client communication, pricing by value, and shipping reliable code. It focuses on: Academic coursework teaches algorithms and theoretical computer science, but leaves students unprepared for client ambiguity, production deployments, and commercial responsibili…

A retrospective on transitioning from computer science theory to delivering production software for clients: client communication, pricing by value, and shipping reliable code.

Who this is for

Engineering students, early-career developers, and freelancers looking for an honest account of learning to build, deliver, and communicate real software for real clients.

Problem

Academic coursework teaches algorithms and theoretical computer science, but leaves students unprepared for client ambiguity, production deployments, and commercial responsibility.

Business outcome

Practical principles for building commercial credibility: scoping projects realistically, communicating trade-offs to non-technical stakeholders, and shipping maintainable software.

Metrics

  • Transited from student coursework to delivering production systems
  • Maintained continuous commercial client delivery across web, backend, and mobile
  • Engineered software deployed and used by real businesses in Kenya and internationally

In university lecture halls, software engineering often feels clean and orderly. Algorithms have definite big-O complexities, textbook problems have clean sample inputs, and exams test whether you memorized the difference between quicksort and mergesort.

Then you land your first commercial client.

Suddenly, nobody is asking you to invert a binary tree. They want to know why their M-Pesa payments aren't updating their accounting records, why the mobile app crashes on older Android phones, or how to launch an operations portal before the end of the month.

This article is an honest reflection on my journey since 2021—from studying computer science fundamentals to engineering production systems as an independent software engineer and professional developer.

The Gap Between Academia and Production

The greatest shock when stepping into the commercial software landscape is discovering that **writing code is only about 40% of the job**. The rest is: 1. **Understanding the underlying business problem**: Often what a client *asks* for is not what they actually *need*. If a client asks for a complex AI chatbot, they might actually need a searchable FAQ and a clear contact phone number. 2. **Defensive architecture**: In school, nobody teaches you what happens when Safaricom's webhook arrives twice with a 30-second delay, or when a user inputs an emoji in a phone number field. 3. **Communication and expectations**: Setting boundaries, communicating technical trade-offs in plain English, and providing realistic estimates.

The Three Rules That Transformed My Engineering

Over the past four years, three principles have guided my growth:

### 1. Choose Boring Technology for Critical Paths When you are a student, it is tempting to use the newest library released on Hacker News last week. But when real money or operational data is on the line, stability is paramount. - Use proven databases like PostgreSQL. - Write clear, idiomatic code before reaching for clever metaprogramming. - Use battle-tested patterns like modular monoliths and clean relational schemas.

### 2. Explain Why, Not Just What Clients and non-technical stakeholders rarely care about your ORM or whether you used Docker. They care about: - Will this system scale during peak hours? - Is our data secure from unauthorized eyes? - How easily can another engineer maintain this system in two years?

Learning to frame architectural choices around **business outcomes and risk mitigation** changed how clients perceived my value. I wasn't just a coder; I was a partner solving real business problems.

### 3. Take Ownership of the Edge Cases Amateur software works when everything goes right. Production software works when everything goes wrong. - What if the network drops mid-transaction? - What if an upstream provider times out? - What if two users edit the same record at the exact same moment?

Designing with failure in mind—implementing idempotent endpoints, transactional outboxes, and proper database constraints—is what distinguishes a true software engineer from a tutorial follower.

Looking Forward

Today, whether I am building a Go modular monolith for internal operations, an Android application for utility customers, or a real-time collaborative workspace, I bring the same discipline: understand the problem deeply, choose the right architecture, and ship software that works reliably under real-world conditions.

Read more about my background, inspect my [production case studies](/case-studies), or [reach out to collaborate](/contact).

Trade-offs

Learned when to choose boring, reliable technology over shiny new frameworks to protect client budgets and timeline guarantees.

Challenges faced

Translating vague client feature requests into concrete technical specifications with verifiable delivery boundaries.