Software Engineer at Bilby Labs--:--

About Ben

Now at Bilby Labs

Ben Alvaro

Software Engineer · Machine Learning

I specialize in creating, developing and maintaining a variety of software and tools, including Machine Learning and AI applications.

Resume GitHub LinkedIn
Currently
  • Software Engineer, Bilby Labs
  • Incoming SDE Intern, Amazon Web Services
  • IT Officer, UniMelb CISSA
  • B.Sc. Computing, University of Melbourne
Melbourne, VICPress ⌘K to find anything

Projects

MUR M26 Race Car

Melbourne University Racing · finished and driving

Working with embedded systems for the first time, I led a team of 7 to design, implement and test software for the MUR M26 race car. I was responsible for the ECU logic, as well as the final integration of all components into the monorepo.

Highlights

  • The M26 is finished and driving.
  • Firmware and a wireless telemetry service for an electric Formula SAE car.
  • Integrated the CAN buses across the car, working with the other sub-teams.
  • Built for Formula SAE-A at Tailem Bend, South Australia.

Built with

C · Zephyr RTOS · Python · Embedded Systems
View the team ↗

Basketball Predictor

Machine learning · live API

An ML based basketball outcome predictor that leverages advanced metrics to forecast NBA game results and identify betting value by comparing fair and market odds.

Highlights

  • Two win-probability models side by side: logistic and normal.
  • Converts probabilities to fair odds and flags games where the market is longer.
  • A GitHub Actions job runs the model three times a day and publishes the picks to Predictor.app.

Built with

Python · nba_api · GitHub Actions · JSON
Open Predictor.appSource ↗

SwiftUI Gym Tracker

Native iOS app

A native iOS application focused on simple UI for tracking user's gym progress. Designed with experienced lifters in mind, the app features minimalist UI and a depth of analysis to help users track their progress and achieve their fitness goals.

Screenshots

Built with

SwiftUI · SwiftData · Swift Charts
Source ↗

CISSA Web Application

UniMelb CISSA · built as IT Director

A full-stack web app for CISSA (Computing and Information Systems Student Association) to display events, engage the community and offer job opportunities.

Highlights

  • Released the new club website and the internal infrastructure behind it.
  • Interactive pages and plug-ins in Python, React and Node.js.
  • Supports events totalling 1000+ participants a year.

Built with

React · Node.js · Tailwind · Python
cissa.org.au ↗

AI Snake Player

Reinforcement learning

Implemented a Deep Q-Network (DQN) reinforcement learning agent trained to achieve optimal pathing and maximum score in the classic game Snake, demonstrating strong foundational knowledge of RL concepts.

Highlights

  • An 11-feature state (danger, heading, food direction) into an 11 → 256 → 3 network.
  • Also trained agents for other games, including Space Invaders.
  • The trained Snake agent runs live in this site as Snake.app.

Built with

Python · PyTorch · PyGame · Reinforcement Learning
Source ↗

F1 Analytics Platform

Data analysis · April 2026

A platform to analyse how lap times change over time in relation to other factors. I used unsupervised ML techniques such as anomaly detection and association rules to find what influences lap times, and to isolate those effects.

Built with

Python · Pandas · Matplotlib · Unsupervised ML
6 items↑ ↓ to browse