Software Engineer · AI · Full-Stack · Automation · Systems · Software Engineer · AI · Full-Stack · Automation · Systems · 
Available for opportunities

Caige

Middaugh

Forward Deployed Software Engineer

Building production systems at the intersection of AI, automation, and full-stack engineering. Currently pursuing an MSAI at UT Austin.

View Projects Get in Touch
caige@portfolio ~
$ caige --help
Available commands: --skills, --projects, --contact, --whoami
$
Scroll
02

About

I'm a software engineer who cares about building things people actually use — and making the people around me more effective. Whether that's shipping production systems, building internal tools that improve a team's quality of life, or creating something a community asked for, the throughline is the same: listen, build, ship, iterate.

At Worlds, I built an internal SDK to streamline how we developed against the platform — centralizing integration patterns and reducing repetitive setup work. At Ethos Group, I was promoted to the R&D team for rapid delivery and led the integration of AI/NLP features across the product suite.

Outside of work, I build for communities I'm part of. AzerothHub started because the WoW community I play in needed a tool nobody had built — now it has thousands of weekly users. PickleBrackit exists because the tournament app our pickleball group used went behind a paywall and organizers needed a free replacement. I listen to what people need and build it.

I'm pursuing an MS in AI at UT Austin while working full-time, and I bring that same energy to everything — collaboration, community, and shipping things that matter.

Currently exploring

LLM-native application architecture, agentic workflows with LangGraph + N8N, and the product surface area between AI APIs and real business workflows.

5+
Years Engineering
3+
Live Side Projects
10+
Projects Shipped to Production
∞
Pickleball Rating
03

Projects

01
Agentic Project Management
A cross-platform desktop app that orchestrates autonomous AI agents to plan, decompose, and execute software development work. Features a 6-step planning flow, Kanban board, multi-role agents, and real-time execution monitoring — built with React, Tauri (Rust), and Claude CLI.
React Tauri / Rust SQLite Claude CLI
View →
02
AzerothHub
A web-based map and data tool for the Bronzebeard WoW server community with thousands of weekly active users. Shows Worldforge items, Mystic Enchants, and rare spawns — integrating with the open-source LootCollector addon. Built from community demand.
Next.js TypeScript PostgreSQL LootCollector
View →
03
PickleBrackit
A real-time tournament management platform for pickleball communities. Round robins, brackets, live scoring, and standings — built to replace a paid tool our community depended on. Mobile-first web app with QR code registration.
Next.js Convex Clerk Tailwind
View →
04
Import Preflight
A B2B SaaS tool that scans Shopify CSV imports in seconds to catch errors before hours-long imports fail. 50+ parallel detectors with semantic understanding of product structure, an algorithmic auto-fix engine, and a Clean Data Score. Built for agencies and merchants.
Next.js Supabase TypeScript Zustand
View →
04

Experience

2025 — Present
Worlds
Remote
Forward Deployed Software Engineer
  • Sole technical owner of large-scale distributed systems processing thousands of events per second across dozens of nationwide sites
  • Architected production-scale CV and ML pipelines, platform SDKs, and agentic workflow layers with direct client-facing accountability
  • Built observability tooling, analytics pipelines, and CI/CD infrastructure via GitHub Actions
  • Designed and deployed 24/7 mission-critical systems — scoping value, negotiating timelines, and delivering production systems end-to-end
Key Projects
Vehicle Gate Automation NDA
24/7 mission-critical system deployed across dozens of client sites. Sole technical architect, developer, QA, and DevOps.
PythonGraphQL APISDKCV / OCR
Factory Tracking & Optimization NDA
Data review dashboards and throughput optimization for manufacturing clients.
PythonGraphQL APIStreamlitN8N
CV Safety Detection NDA
Person down, red zone entry, collision path detection. Data review and optimization pipeline.
PythonGraphQL APIN8NCV
MCP Server & Data Classification NDA
MCP server giving Claude access to self-hosted N8N and product data. Classifies data (true/false positive) to optimize workflows.
MCPClaudeGraphQL APIN8N
Internal SDK NDA
Dev tools streamlining platform integration — centralized access for FDEs and sales to build faster on the product.
PythonGraphQL APISDK Design
N8N Platform for Clients NDA
Dozens of custom N8N nodes on self-hosted platform, handed off to consulting firms and clients to build on.
N8NGraphQL APICustom Nodes
2021 — 2025
Ethos Group
Remote
Full Stack Software Engineer
  • Built and maintained large-scale distributed infrastructure (IoT Edge) serving hundreds of client locations across the U.S. (Azure IoT Edge, Docker, Kubernetes)
  • Promoted to R&D team for rapid delivery of high-priority products, owning initiatives from concept through production
  • Spearheaded integration of NLP models and LLMs into production products, plus prompt engineering, token optimization, and real-time monitoring of non-deterministic AI behavior
  • Engineered data transfer compiler integrating internal systems with 12+ external partner APIs using scalable microservice architecture
Key Projects
Large Data Compiler NDA
Translated data from 12 external partners to internal teams with scalable microservice architecture.
.NETAngularSQLEntity Framework
Sales Replay NDA
Record, transcribe, and apply AI features — semantic analysis, Q&A extraction, rankings against training manual, AI feedback. Microservices synced with Playbook.
OpenAI.NETAngularSQLIoT Edge
Playbook NDA
Knowledge base with scripted homework, recording, and AI roleplay for sales training. RAG architecture for AI responses.
.NETAngularPythonOpenAIRAG
2025 — Present
UT Austin
Graduate
MS in Artificial Intelligence (In Progress)
  • Pursuing MSAI alongside full-time engineering work
  • Research focus on trajectory-based pedestrian-vehicle collision risk prediction using ML
  • Applying coursework directly to production AI systems and agentic workflow design
05

Skills

Let's work together

Let's
build
something.

Open to senior engineering roles, founding engineer opportunities, and teams where technical ownership and client impact are valued.

caige.middaugh@hotmail.com LinkedIn ↗ GitHub ↗