Information management · Data governance · Analytics · Reliable automation
I turn complex systems into trusted ones. My background spans more than 10 years across production operations, database operations, digital asset management, reporting, metadata, taxonomy, and process improvement. The work here shows how I translate that experience into tested tools and analyses with explicit controls, documented results, and clear limits.
M.S. in Information Management · Certificate in Digital Curation · Dominican University
Current focus: Building local-first tools and data systems where evidence, rollback, and clear operator control are part of the product.
Each figure is defined in the public project summaries and source ledger below.
| Evidence | What it means | Why it matters |
|---|---|---|
| 13,333 public inspection records across three Chicago ZIP codes | One row is an inspection, not a restaurant; the study covers 2010–2018 | Shows reproducible analysis and careful separation of inspection outcomes from facility-risk classification |
| 36 authored cases across six shifts | Every case participates in branching state and a full-campaign acceptance test | Shows product scope, persistent state, accessible interaction, and end-to-end testing |
| 17 public projects reconciled to exact GitHub source | Each ledger entry is tied to the version and reviewed default-branch commit actually present | Shows portfolio-wide version governance and avoids presenting older code as a newer build |
I’m open to senior roles in analytics, data governance, information management, and reliable automation. My strongest work is grouped below into flagship projects, a delivery framework, and broader applied projects.
- BotOps Manager — Verifies launchers and process ownership before control, consolidates health evidence, and creates limited, redacted support exports for independent Windows automation.
- Digital Asset Governance Audit — Turns schema health, provenance coverage, confidence signals, and duplicate candidates into a prioritized stewardship queue with clear reasons and next steps using synthetic data.
- MediaTaggerBot — Uses evidence-based matching, complete-scan gates, dry-run review, recorded changes, result verification, verified renames, and rollback manifests. Uncertain or preservation-sensitive files remain visible for review instead of being forced into a match.
- Reliable Project Delivery Framework — Turns conflicting sources, machine differences, risky changes, and incomplete evidence into a portable release path with clear rollback and independent local operation.
- Chicago Food Inspection Outcomes, 2010–2018 — Reproducibly analyzes 13,333 public inspection records, separates outcome from facility risk, and surfaces decision-relevant follow-up questions without overstating the data.
- Avalon Q Supervisor — Turns local device telemetry into clear health states, requires separate configuration and runtime gates before recovery, enforces durable restart limits, and records privacy-conscious decision evidence.
- Automation Reliability Case Studies — Three documented studies show how outcome checks, identity-bound supervision, integrity verification, limited recovery, and explicit stop conditions contain risk across uncertain system boundaries.
- Beta Earth — A working local-first Python RPG with portable launchers, resilient saves, release validation, and verified public source at v0.4.11. The newer recorded v0.5.0 final remains visible behind an exact-source checksum gate instead of being falsely represented by the older tree.
Additional tools and learning releases
- Safe Video Downloader — Retrieves authorized media through shared desktop and command-line planning, public-network checks, duplicate controls, isolated workers, fallback media-signature validation, accurate status for mixed-result batches, and verified output.
- MP3 Downloader — Turns one authorized URL into a validated MP3 through metadata preflight, retry limits, SHA-256 and SQLite duplicate reconciliation, visible output, and tightly redacted support exports.
- Image Downloader — The 2026.08.02.1 release collects permitted images with controlled discovery, a five-second per-image network budget, credential-safe evidence, Windows-safe filenames, resumable state, duplicate controls, and atomic finalization.
- Large Text Chunker — Splits local documents at readable boundaries, verifies exact reconstruction, and records SHA-256 and token evidence. Version 1.10 adds offline estimate, optional exact, and automatic fallback modes plus a dated upload-cap advisory.
- NetLossDoctor — Correlates gateway, route, DNS, TCP, latency, Windows event, and optional trace evidence into time-boxed, redacted support packages without changing network configuration.
- LAN Router Comms — Provides cloud-free authenticated text and resumable file exchange between trusted Windows PCs using durable queues, replay protection, session limits, and verified delivery receipts.
- Windows Health Audit — Creates a read-only, non-elevated Windows health baseline with failure-isolated collectors, privacy-preserving normalization, transparent review thresholds, and no network or system changes.
- Inbox From Hell — A six-shift workplace-horror browser game with 36 branching support cases, permanent department progression, accessible controls, resilient local saves, and automated full-campaign acceptance testing.
- Kalshi 10×1¢ Public Edition — A checksum-controlled educational entry planner demonstrating explicit activation, authenticated API use, duplicate-action prevention, execution caps, and safe stopping rules.
- Kalshi 15-Minute Sell Preview — The downloadable v41.22.3 public preview is a checksum-consistent dry-run exit planner with live writes blocked, bounded polling, state reconciliation, release verification, and clear limits on what the build can do.
Public project release reconciliation records the version actually present in each repository, the latest verified build, the exact reviewed default-branch head, and any checksum-gated promotion blocker.
Public dependency reconciliation records dependency modes, exact pins and hashes, installation and consistency checks, deferred major-version reviews, and private-workspace boundaries across all public repositories.
- Keep work local when privacy or operator control matters
- Preview potentially destructive changes before applying them
- Verify process identity and ownership before process control
- Prefer structured evidence and checksums to assumptions
- Keep scope, thresholds, and operating limits in one reviewable configuration
- Stop safely when recovery cannot be verified
Fixtures in these repositories are synthetic or non-sensitive; credentials, runtime data, generated diagnostics, and downloaded media are not committed.