Software Developer at APS Data Technologies, building production AI-powered platforms and high-load B2B systems. Studying Computer Science with a Business minor at Michigan State University — GPA 3.8, graduating May 2026.
Focused on distributed systems, AI/ML, and high-performance computing. I build things that ship — from enterprise platforms processing thousands of applications daily to hackathon projects that turn into real products.
Career
Experience
April 2024 — Present
Software Developer
APS Data Technologies — EdTech B2B
Building production AI-powered ATS systems processing 1,000+ applications daily for K-12 school districts. Full-stack work across Django REST, Next.js, PostgreSQL, and AWS. Led development of a RAG-powered college application platform with OCR document pipelines and vector search.
September 2024 — December 2024
Software Engineer
Safety Straw — Seed-stage Safety Tech Startup
Delivered e-commerce platform with 4-person team 25% ahead of timeline. Built frontend and CI/CD pipeline reducing deployment from hours to minutes. Architected RESTful API connecting React to Express/MongoDB with JWT auth and role-based access control.
September 2024 — May 2025
Operations & Growth Lead
Imagine Software — Student Organization
Scaled student tech org 55% from 450 to 700 members. Conducted 20+ technical interviews for client project staffing.
2022 — May 2026
B.S. Computer Science
Michigan State University — Business Minor
GPA 3.8. SpartaHack XI Blockchain Track 3rd Place.
Case Studies
Selected Work
APS Hiring Platform
2024 — Present
Production ATS processing 1,000+ applications daily for K-12 school districts.
Full hiring pipeline: job posting, applicant tracking, interview scheduling, offer generation, document collection, and onboarding — all in one system.
Django 5.2 + DRF backend with PostgreSQL, serving a Next.js 15 + React 19 frontend. Type-safe API layer with TypeScript throughout.
Multi-tenant architecture supporting concurrent school districts with isolated data. Role-based access across administrators, HR staff, and hiring managers.
DocuSign integration for electronic signatures, SSO for district-wide authentication, and automated email workflows for candidate communication.
Public demo represents ~40% of the production system — proprietary integrations and multi-tenant logic removed.
RAG-based AI chat assistant: OpenAI embeddings stored in Qdrant vector database, with retrieval-augmented generation for contextual answers about universities and programs.
OCR pipeline for automated transcript and resume parsing — extracts structured data from uploaded documents for application processing.
AWS infrastructure: S3 with pre-signed URLs for secure file management, EC2 for compute, RDS for managed PostgreSQL.
Django REST + Next.js 15 full-stack architecture. Docker Compose for local development and deployment orchestration.
Demo represents ~30-40% of the production system. Full version includes additional AI features and institutional integrations under NDA.
GPT-4 generates unique coding exercises with test cases across React, JavaScript, Python, and C++. Structured output ensures consistent exercise format.
Monaco Editor integration (same editor as VS Code) with language-specific syntax highlighting, autocompletion, and split-panel UI.
Sandboxed code execution workers: user-submitted code runs in isolated environments with resource limits. Real-time test validation against generated test cases.
FastAPI + PostgreSQL backend. Zustand for frontend state management. Docker Compose for reproducible multi-service development.
Full exercise lifecycle: generate → code → execute → validate → feedback. All in-browser, no local toolchain required.
Stride scheduling algorithm: processes have a virtual time position (pass) and stride inversely proportional to priority. Scheduler always picks the process furthest behind in virtual time — guaranteeing fair, proportional CPU allocation.
Tiered execution system with 4 importance levels: CRITICAL (40% frame budget), GAMEPLAY (30%), ECONOMY (20%), COSMETIC (10%). Each tier backed by its own Scheduler instance via composition.
Soft budget enforcement for CRITICAL tier (up to 250ms overage allowed), hard budget cutoff for all other tiers. Prevents cosmetic processes from starving gameplay-critical ones.
C++23 compiled to WebAssembly via Emscripten. Designed interfaces between world state, agent behaviors, and rendering systems across distributed development teams.
Full test suite: unit tests for base Scheduler (add/remove/priority/peek) and TieredScheduler (budget enforcement, tier isolation, edge cases). Assert-to-throw conversion for robust error handling.
