An overhead shot of a laptop, notebook, and coffee cup at midnight — franzdomingo.tech

Third Year as a Software Engineering Student: The Year Everything Got Real

If first year taught me how to write code, and second year taught me how to use tools —
Third year taught me how to build under pressure.

This was the year where concepts met deadlines.
Where theories collided with power outages.
Where code wasn't just "run and done" — it had to be reliable, scalable, and defendable.

🧩 The Curriculum That Tried to Break Me

Third year went beyond syntax. It gave me real engineering context.

Core Subjects I Wrestled With:

Mobile Programming: My gateway to Flutter, Dart, Swift — and the start of the LiftTrack mobile app.
Network & Communications 2: Concepts like routing, port forwarding, socket connections — essential for WebSocket communication in LiftTrack.
Software Engineering 1 & 2: Structured our project planning and helped us design before we built.
Modeling and Simulation: Made me think of systems as data-driven representations — useful for posture analysis in our AI project.
Image Processing: Helped understand how frames were captured and transformed in the 3DCNN model.
Parallel and Distributed Computing: Reinforced our struggles with compute limits — especially training the AI model under constrained hardware.
Programming Languages & Number Theory: Gave theoretical depth to language design, parsing, and optimization.
CS Project 1: The formal stage where LiftTrack lived or died.

Also: Project Management, Business Process, and Technopreneurship shaped how I thought about the why, not just the how of building.

⚙️ LiftTrack: The Overkill Thesis (Revisited)

A mobile app powered by a 3D Convolutional Neural Network (3DCNN) to analyze weightlifting posture in real-time.

Sounds exciting. It was — until it nearly broke us.

Why We Failed (Twice):

Platform lock-in: Apple’s iOS debugging restrictions delayed testing.
Compute bottlenecks: One teammate’s machine died training the AI model.
WebSocket failure: Our ISP blocked ports critical to our real-time sync.

Why We Eventually Passed:

Switched to cloud hosting on Google Cloud.
Built a Flutter app with Dart and Swift for full cross-platform deployment.
Used GitHub Actions for deployment and CI/CD pipelines.
Shifted to Firebase (basic) to store form data and survey feedback when MySQL integration lagged.
We restructured the app and even mocked the WebSocket layer locally to meet our defense deadline.

💻 Tools & Stack I Used in Real Production

Languages

Python: For training and inference of the 3DCNN
Dart + Swift: For mobile dev in Flutter and native iOS fixes
SQL + Firebase: Data storage for surveys, model outputs
JavaScript (ES6+): For static blog rendering and portfolio logic

Frameworks & Platforms

Flutter (Dart): Cross-platform mobile app development
Flask: Python backend for model inference
Firebase: Lightweight, fast integration for survey collection
GitHub Actions: CI/CD and automation
Google Cloud Platform (basic): Hosted model and inference endpoints
Vercel: Hosted portfolio with .mdx blog content

IDEs & Tools

VS Code, Android Studio, Xcode
Git + GitHub for full version control and branching
Canva for visuals and project documentation

🧠 Academic Concepts Applied in Real Projects

Image Processing + Simulation → Used to preprocess lifting videos for model input
Parallel Computing → Helped us understand model threading (though we couldn’t use it fully)
Project Management → Allowed us to break down LiftTrack into milestones to avoid scope creep
Networks → Directly used in handling WebSocket timeouts, port binding, and local IP testing
Programming Languages Theory → Showed me how Dart handles compilation vs JavaScript

📉 What Nearly Broke Me

TTL and DNS record delays killing a late-night deployment
React Native vs Flutter decision spirals
Debugging a Swift iOS .ipa export alone (I handled this version solo)
The moment our backend and frontend finally worked… and then the ISP blocked our port

🧠 What I Gained

Real-world confidence using Python for AI tasks and servers
Production-ready habits: CI/CD, logging, backups, versioning
Mobile deployment knowledge (Dart, iOS, Firebase workflows)
Scoping instincts: how to shrink ambition into something that ships
Understanding that failure isn’t the opposite of success — it’s the path to it

“Third year didn’t just challenge me. It redefined me.”

🚀 Looking Ahead

After the thesis defense:

I started studying Docker for lightweight local deployment and consistent testing
Began exploring AWS as a scalable alternative to Firebase and GCP
Took an interest in Cybersecurity (inspired by our WebSocket vulnerabilities)
Started building reusable prompt templates for AI-powered development tools