Ty Rimedio

Data science, machine learning, and predictive systems

Ty Rimedio

I build machine learning pipelines, backtesting systems, and analytics products that still hold up after the prototype phase.

I'm a computer science senior at Indiana University focused on turning messy, real-world data into models and systems people can actually trust, use, and ship.

Selected workReach outDownload resume

Discipline

Machine learning, data science, and applied systems work.

Goal

Be memorable enough that someone reaches out or forwards my name.

About

Built around technical depth, not portfolio theater.

The work I care about most sits in the uncomfortable middle ground between modeling, engineering, and decision-making.

I'm a computer science senior at Indiana University focused on data science and machine learning, especially when the work has to survive real constraints and real users.

Right now I'm running an NBA prediction pipeline on a Raspberry Pi, iterating on feature engineering, backtesting, and evaluation so the output is useful outside a notebook.

Focus
Data Science & ML
University
Indiana University
Graduation
May 2026
Location
Bloomington, IN

I'm most interested in roles where clean pipelines, reproducible experiments, and measurable model performance matter as much as the model itself.

Projects

Work that holds up outside the prototype.

The throughline is simple: build from real data, validate under real constraints, and make the result usable.

Featured project

Project Lead / Mar 2026 - Present

NBA Prediction System

End-to-end sports analytics system that ingests NBA and odds data, engineers leak-aware features, and generates daily game predictions — evaluated on held-out data, not cherry-picked results.

  • 67.1% logistic accuracy across 1,075 held-out games with documented log loss and Brier score
  • Benchmarked against Pinnacle closing lines — approaches market-level accuracy, no claimed edge
  • Runs autonomously on a Raspberry Pi with four daily slate refreshes and lock-based scheduling

Python / scikit-learn / LightGBM / pandas / NumPy

Live Dashboard

Calibration

1,075 held-out games

25%50%75%25%50%75%PREDICTEDACTUALperfect

67.1%

Accuracy

0.6065

Log loss

0.2101

Brier

Monthly accuracy, Oct – Mar

OctNovDecJanFebMar

Additional work

Product work, simulation work, and lower-level implementation detail.

Project Lead

Momentum Fitness

May 2025 - Present

Personalized fitness and nutrition tracking system built around recommendation logic, user data, and progress analytics on iOS.

  • Structured workout and nutrition logging with USDA API integration and macro trend tracking
  • Personalized coaching workflow driven by user goals, habits, and logged activity
  • Interactive progress charts, Firebase auth, and a home screen widget for daily feedback

Swift / SwiftUI / Firebase / USDA API

Contributor

Multi-Robot Pathfinding

Mar - May 2025

Algorithmic simulation project focused on multi-agent path optimization, real-time collision avoidance, and replayable systems analysis.

  • A* algorithm for optimal path generation across multiple agents
  • Priority-based collision avoidance with dynamic replanning
  • Replay system with movement visualization and congestion hotspot analysis

Python / Pygame / A*

Watch Demo

Project Lead

Image Editor

Nov - Dec 2024

Image processing project centered on pixel-level transformations, composable effects, and interactive experimentation.

  • Grayscale, color manipulation, rotation, mirroring, and pixelation effects
  • Undo/redo system and file I/O with object-oriented design

Java / Swing

Watch Demo

Capabilities

A toolkit built for applied machine learning work.

My strongest work sits at the intersection of modeling, validation, and the infrastructure required to keep technical ideas useful.

Languages

Python / SQL / Java

ML / Data

pandas / NumPy / scikit-learn / LightGBM / SHAP / Feature Engineering / Backtesting

Tools & Infra

Git / systemd / Workflow Automation / Raspberry Pi / ML Pipelines

Concepts

Machine Learning / Predictive Modeling / Data Analysis / Model Validation / Data Visualization

Contact

If you're building serious data products, I'd like to hear about it.

I'm looking for opportunities in data science, machine learning, and data engineering. If my work looks relevant, reach out directly or pass my name along.

The bar is simple: specific enough to earn a conversation, and credible enough to be worth forwarding.

trimedio03@gmail.comEmailgithub.com/tyrimedioGitHublinkedin.com/in/ty-rimedioLinkedIn

Ty Rimedio · 2026