Chase
Key
AI Engineer with roots in professional aviation and enterprise data systems. I design and ship production AI — from autonomous agents and local LLM infrastructure to real-time game servers and client deployments. My open-source work has been submitted to Anthropic's skills repository.
About Me
I am genuinely grateful for every opportunity that has led here. My path has not been linear — aviation, enterprise data analysis, and now AI engineering — but each chapter built something the next one needed.
Aviation gave me systems thinking and a deep respect for what happens when precision fails. The FAR-AIM — a massive cross-referenced regulatory framework governing all U.S. flight operations — became my earliest model for information architecture. That principle carries into everything I build.
Data analysis grounded that thinking in real-world complexity: multi-year anomalies, cross-jurisdictional investigations, SLA risk, and the weight of getting it right for teams depending on accurate data.
AI development gave me a new medium to build things that genuinely help people work better. I believe the most valuable thing an engineer can do is understand a system well enough to make it trustworthy — whether that system is a flight regime, a data pipeline, or an autonomous agent.
B.S. Aviation Sciences, 2010
Google Professional Data Analytics Certificate
Technical Skills
AI & Agentic Systems
Data Engineering & SQL
Software & Infrastructure
Projects
Designed and published a modular, open-source cognitive AI architecture spanning six interoperable repositories. Each module is independently deployable and composable — submitted for inclusion in Anthropic's public skills library, reviewed against their contribution standards for structure, documentation, and reusability.
| aurelion-kernel-lite | 5-layer cognitive structure templates for organizing complex reasoning |
| aurelion-memory-lite | File-based persistent knowledge graph (Python) |
| aurelion-advisor-lite | Strategic planning templates and methodology library |
| aurelion-agent-lite | 100+ AI collaboration prompts and agentic thinking protocols |
| aurelion-nexus-lite | Story-agnostic NPC and world simulation framework (Python) |
| aurelion-hub | Central orchestration hub and documentation index for the suite |
All six repos are also packaged as pip-installable Python modules under aurelion-eco/modules/ — composable, editable packages for any project.
Architected, built, containerized, and deployed a fully-functional autonomous AI agent with persistent memory, contextual reasoning, and semantic lore retrieval (ChromaDB). Concept to production in under five hours.
- Lazy indexing: cold-start reduced from 30s to <2s
- Three-tier character resolution fallback for zero-downtime operation
- Cross-platform encoding hardened (Windows dev → Linux container)
- GPT-4 → GPT-3.5-turbo cost routing: ~$0.02–$0.05/query
Five-layer knowledge architecture: 35+ interconnected documents, 15,000+ lines of structured content, Python-powered search library, and semantic knowledge graph (JSON). Modeled on FAR-AIM indexing principles. Accepted for organizational pilot across a 17-analyst team.
- 4-week peer-to-peer training curriculum (team-adopted)
- Multi-stage investigation decision tree (validated in production)
- Enterprise data governance framework — 10 sections
- Zero-cost knowledge system deployed without additional budget
Autonomous compliance and data governance engine that audits datasets against structured regulatory profiles (GDPR, CCPA, HIPAA). Schema-driven rule evaluation — no hardcoded logic.
- Profile-based checks: required fields, data types, format validation
- Severity scoring: CRITICAL / HIGH / MEDIUM findings with audit trail
- FastAPI web dashboard with file upload and live audit results
- Designed as open-core — profiles are community-extensible
Real-time testing and observability console for multi-model AI agents across Groq, Gemini, OpenAI, Claude, Mistral, and Ollama. Unified interface with trace logging and workflow execution.
- 8 live agents: Groq, Gemini, Ollama, OpenAI, Claude, Mistral, Supervisor, Unified
- SQLite trace store — every run logged with input, output, latency, and model
- Workflow tab: multi-step agent chaining with configurable steps
- Memory tab: persistent context injection across agent runs
Fully self-hosted agentic AI running on an external drive. Ollama-powered with custom om-deep and om-fast models — GPU maxed at 16,384 context length. 12,406 documents indexed across seven ChromaDB collections. MCP server bridge enables live tool injection into any connected agent.
- 12,406 docs across 7 ChromaDB collections: knowledge, stonecrest, research, facts, capabilities, projects, preferences
- Custom Ollama models (om-deep / om-fast) — 7.9 GB GPU, 16,384 ctx, 100% utilized
- MCP stdio server — 7 tools: om_chat, om_research, om_remember, om_retrieve, om_ingest_url, om_health, om_inspect_self
- 15 active capabilities (14 stable, 1 experimental) — inbox watching active
Orchestration gateway bridging GitHub Copilot (Claude Sonnet 4.6, medium reasoning) to Om's MCP tool server and a live Telegram bot. Persona-driven with a SOUL.md character file that hot-reloads without restart. MCP tools inject at session startup — 7 Om tools available to any connected agent.
- Telegram gateway: CK_Master_Bot_1 — live conversational AI accessible on mobile
- 7 Om MCP tools injected at session startup via stdio bridge
- SOUL.md persona (King Aurelion "Rell" Valenhart) — hot-reload, no restart required
- GitHub Copilot as reasoning brain — Claude Sonnet 4.6, medium reasoning budget
Playable text-based multiplayer RPG set in the Stonecrest world. Players create characters, choose classes, fight enemies, and progress through zones. Real-time WebSocket combat server deployed on Railway.
- Character creation with class system — persistent SQLite game state
- Zone-based progression with seeded enemies and encounters
- Real-time WebSocket server for live multiplayer sessions
- Companion to Memoria Engine — shares the same Stonecrest world lore
Experience
- State Subject Matter Expert for four U.S. states — responsible for data quality, feed health, SLA compliance, and operational continuity across departments
- Resolved a multi-year data volume anomaly in 8 days through systematic SQL investigation across 3M+ records; engineering team confirmed root cause and deployed fix, restoring accuracy across multiple jurisdictions
- Delivered 30%+ query performance improvements through execution plan analysis and indexing strategy; optimization patterns adopted as team standards
- Reduced team manual processing workload by 20%+ through Python automation scripts
- Built and maintained 3+ Tableau dashboards for live feed health monitoring and SLA risk tracking
- Proactively designed and deployed an enterprise-grade data governance framework (10 sections: data sourcing, PII compliance, lifecycle management, emergency procedures, audit trail)
- Contributed to resolution of a critical QA tooling outage through rapid system-level diagnosis; provided immediate workaround that restored team productivity within minutes
- Completed mandatory Information Security training 24 days ahead of organizational deadline; integrated requirements into team governance documentation
- Identified and proposed a zero-cost team knowledge management system using existing tooling; leadership approved for organizational pilot
- Designed and delivered a formal 5-week analyst onboarding program including structured curriculum, hands-on tool training, and a standardized QA competency assessment
- Mentored 14 analysts through complex investigations, technical troubleshooting, and cross-team research methodologies
- Facilitated 9.5+ hours of formal training delivery; all participants transitioned to independent assignments upon completion
- Converted undocumented operational processes that existed only as tribal knowledge into reusable, versioned SOPs
- Promoted from contractor to full-time employee based on demonstrated technical performance and initiative
- Self-directed learning of SQL and Python applied directly to production workflow automation and data quality improvement
- Developed reference documentation and QA methodology guides later incorporated into team-wide SOPs
- Advanced from initial onboarding to independently conducting multi-jurisdictional investigations within the first year
Education & Certifications
B.S. · Professional Pilot Aviation Sciences
Aviation training established a foundational principle applied to every engineering project since: in safety-critical systems, navigability and clarity of information architecture are not optional — they are the system. The FAR-AIM became the direct architectural model for every knowledge framework built since.
Certifications
-
CS50p: Introduction to Programming with Python Harvard University / edX · Oct 2022
-
Google Data Analytics Certificate Google / Coursera · Aug 2023
-
Python for Data Science Coursera · 2024
-
Querying SQL Databases: Learning SQL Using Prompt Engineering Skillsoft · Aug 2025
-
Enterprise Security: AI, Generative AI & Cybersecurity Skillsoft · Feb 2026 · Credential #175329234
-
CompTIA SecurityX: AI Security Challenges Skillsoft · Mar 2026 · Credential #175892826
-
JavaScript — Data Structures & Algorithms freeCodeCamp · Feb 2023
-
Advanced SQL · 60% complete DataCamp · Target: Q3 2026
-
Advanced Python Data Science · 30% complete Coursera / DataCamp · Target: Q4 2026
Contact
Open to AI Engineer, Gen AI Engineer, and senior data roles. I respond to thoughtful messages — feel free to reach out directly.