Portfolio

Applied ML &
systems engineer

Materials-science master's graduate, with minors in Software Development and AI/ML. Interested in increasingly challenging tasks, from real-time systems to data-heavy pipelines to interactive simulations. Comfortable picking up unfamiliar tools when a project demands them.

01

2025

MIMOSA: CNN particle analysis pipeline

Automated particle dispersion analysis software featuring a convolutional neural network for batch analysis of microscopy images. Reduces per-image dispersion calculation from ~5 minutes (with existing tools) to 5 seconds. Developed as part of a master's thesis.

Python TensorFlow Image Processing
View on GitHub
View Master's Thesis

02

2023

CMA: Correlative microscopy analysis tool

Correlative microscopy analysis tool built with another student for our thesis instructors, split across a defined frontend/backend boundary: I owned the MATLAB App Designer GUI (layout, interaction design, styling), while my partner handled the backend image-processing pipeline. Coordinated work on a shared Git repository, partitioning edits and handling merges to avoid conflicts.

MATLAB GUI Collaboration
View information
View Bachelor's Thesis

03

2025

Nappulapeli: Chaotic Party Game

An autobattling party game built for a friend group trip, where players control characters from an Android app while battles play out live on a shared host screen.

  • Optimized pathfinding, rendering, and game state ticking.
  • Custom socket server handling real-time state sync across multiple concurrent clients.
  • AI avatar pipeline (background removal + facial recognition) turning player selfies into in-game character cutouts.
  • Spatial audio engine written from scratch: 100+ simultaneous sound sources with low-pass filtering. Released as a PyPI package.

Host in Python/Pygame, client in Kotlin/Android Studio. Developed over 6 months.

Python Kotlin Sockets Audio Processing
View game on GitHub
View app on GitHub

04

2026

Nappulapeli 2 WIP

A party dungeon crawler with a shared host screen, styled as a nightclub crawl with Slay the Spire-style run progression and shops. Players manage their characters through a locally hosted React web app, while the host screen shows a live combat view. Player and enemy portraits are generated through a local SDXL + IPAdapter + ReActor pipeline for stylized character art. Combat uses a full minimax-based AI to ensure optimal decision-making. Music-synced combat pacing. Host in Pygame. Currently worked on, started June 2026.

Python GenAI Sockets React
coming soon.

Engineer with a background spanning game development and scientific computing. Completed a master's degree with a thesis on applying deep learning to materials analysis, specifically training a CNN to automate what was previously a manual, subjective process.

Drawn to problems that require both rigorous engineering and creative thinking: managing complex overlapping systems, difficult problem solving under constraints, and building intuitive user experiences for complex tools.

Comfortable owning a system end-to-end or working against a shared boundary. CMA was built two-person on a shared repo, coordinating merges across a defined frontend/backend split. Looking for challenging work in applied ML, computer vision, or scientific computing.

Languages

  • Python
  • Kotlin
  • JavaScript

Tools

  • TensorFlow
  • OpenCV
  • NumPy
  • Git

Areas

  • AI/ML
  • Game Development
  • Data processing
  • Computer Vision