Operational Engineering · Systems Architecture

Documenting the
Systemic Shift

I'm Sean Deardorff. I build high-performance operational systems, documenting the transition toward automated enterprise blueprints and consumer-centric AI orchestration. This site serves as an engineering archive for my open-source data pipelines, autonomous agent architectures, and localized utility roadmaps — built with vanilla technologies, optimized for execution speed.

View Engineering Logs Examine Repository
3+
Engineering Logs
10
Repo Assets
0
Vendor Bloat
100%
Native Execution

Operational Documentation

Field Engineering Logs

Technical post-mortems and architecture breakdowns. Every entry identifies the failure state, isolates the root cause, defines the engineering response, and provides a reusable framework for AI automation.

Edge AI Published
Post #001 · July 2026

The NPU Liberation Framework

Forcing ONNX runtime execution onto bare-metal Qualcomm Hexagon hardware under strict 8GB RAM constraints. To bypass closed-source cloud telemetry, I deployed an offline-first edge architecture utilizing Tauri v2 and PocketBase. The entry documents how we eliminated model hallucinations via audio activity detection gates and achieved significant memory footprints reductions.

🕐 6 min read 🏷️ NPU Execution · Tauri v2 · PocketBase
Cloud Native Draft
Post #002 · August 2026

Autonomous Corporate Blueprinting

Designing fully automated corporate operating structures running distributed cloud-native AI agents. This post outlines how Next.js panopticon dashboards deployed on Cloud Run orchestrate edge-harvested data streams to trigger vertex-driven decision loops, shifting operational administrative loads entirely to cloud-native actors.

🕐 8 min read 🏷️ Terraform · Vertex AI · Cloud Run
Client-Side Coming Soon
Post #003 · September 2026

Job Search Intelligence Delta Scoring

Analyzing career trajectories using browser-native score models in IndexedDB. By engineering custom-distilled O*NET/ESCO anchors, we calculate resume matching ($\Delta X$), trajectory alignment ($\Delta Y$), and knockout risk ($\Delta Z$) in real-time, executing localized Transformers.js models with zero cloud egress cost.

🕐 10 min read 🏷️ IndexedDB · Transformers.js · browser-native
⚙️

Root Cause Automation

The temporary patch is never the deliverable. Every entry isolates the structural origins of pipeline failures and agent limitations before deploying production code.

🔧

Vanilla Architecture

All client-side applications run locally with zero external network dependencies — no frameworks, no third-party tracking, no corporate cloud dependencies.

📐

Verifiable Repositories

Every system detailed in the portfolio points directly to real open-source repositories. We build working, inspectable code rather than high-level enterprise concept diagrams.

If Something Here Resonates

These case studies are intentionally detailed. If the methodology connects with an automation problem you're working through — a broken pipeline, an alignment gap, or a client-side utility need — let's start a conversation.

Start a Conversation Review the Repositories