01
↗
Turbofan Predictive Maintenance
An end-to-end prognostics pipeline for turbofan jet engines using Transformer-based time-series modeling, anomaly detection, and Remaining Useful Life prediction.
✦ Software engineer · ML systems · full-stack AI products
AI systems with practical precision.
I build ML-powered software that connects deep learning, reliable backend systems, and useful product experiences—from predictive maintenance pipelines to cross-platform AI assistants.
const engineer = {
name: "William Opyrchal",
focus: ["ML systems", "RAG", "full-stack AI"],
tools: ["PyTorch", "Flask", "Flutter", "Gemini"],
currentlyBuilding: "predictive systems that explain themselves"
};
6+
AI, systems, and embedded projects
20+
technical tools across ML and software
100%
focused on useful, interpretable AI
Selected work
Built at the edge of AI and engineering.
William’s work combines model development, data pipelines, real-time software, retrieval-augmented LLMs, and deployable applications.
02
↗
AskMickey
A cross-platform AI Disney World assistant with RAG, Mixture-of-Experts routing, agentic AI patterns, live attraction data, queue times, and geolocation-aware navigation.
03
↗
Applicant Tracking System
Full-stack ATS components deployed on Amazon EC2, including automated cron jobs that ingest applicants from IMAP email using pattern recognition.
Toolkit
Research depth. Shipping discipline.
Python
PyTorch
Transformers
RAG
Agentic AI
Time-Series Forecasting
Anomaly Detection
Flask
Flutter
Dart
C++
Java
JavaScript
SQL
Linux
Amazon EC2
MySQL
Arduino
Raspberry Pi
Git
Resume
William Opyrchal
Computer Science graduate from the University of Florida focused on machine learning systems, full-stack AI applications, predictive maintenance, and interpretable model-driven products.
william.opyrchal@gmail.com
248-550-3116
LinkedIn
Experience
Machine Learning Software Engineer Intern
Epcom Corporation · Summers 2024 – 2025
Researched and built a Transformer-based predictive maintenance system for multivariate time-series sensor data, integrating retrieval-augmented LLMs with historical maintenance records and failure-mode data.
Software Development Intern
Epcom Corporation · Summers 2020 – 2023
Developed full-stack components of an Applicant Tracking System using Linux, Apache, MySQL, PHP, JavaScript, HTML, and Bootstrap, deployed on Amazon EC2.
Education
University of Florida
B.S. Computer Science · August 2025
College of Liberal Arts and Sciences. Additional independent coursework includes Harvard CS50, MIT OCW mathematics and physics, MIT AI, and Stanford CS229.
Core Strengths
ML Systems
Transformers
RAG
Time Series
Full Stack
Embedded
About
Early career. Unusually broad range.
William is a software engineer with experience spanning machine learning, full-stack development, embedded systems, mobile apps, and applied AI. His strongest work sits where models meet real-world software: pipelines, APIs, retrieval systems, and interfaces that help people make better decisions.
His recent work includes a Transformer-based predictive maintenance system, a NASA CMAPSS-inspired turbofan prognostics pipeline, an LLM-powered Disney World assistant, and earlier production internship work on an Applicant Tracking System deployed on AWS infrastructure.