Skip to content
View RedBeret's full-sized avatar
😀
Studying
😀
Studying

Organizations

@espinoza-FlatironSchoolSE

Block or report RedBeret

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
RedBeret/README.md

Steven Espinoza

Senior Systems Engineer · AI Agent Tooling · Edge & Field Operations

stevenespinoza.dev  ·  LinkedIn  ·  espisteven5@gmail.com


👋 What I build

Senior systems engineer and Army veteran building AI systems that hold up in the real world — offline-first field tools, guardrails for coding agents, and pipelines that handle messy data in constrained environments. Twelve years across IT, network engineering, and full-stack development.

Everything here is independent side work, built on my own time, MIT licensed.


🧭 Forward Deployed Engineering Flagships

Five connected projects built as one field-engineering workflow: realistic source data, measurable cleanup, RAG evaluation, offline delivery and database drift detection.

Project What it proves
acme-parts-cloud One-command synthetic enterprise API, UI, Postgres data, deterministic messiness and manifest-v2 ground truth
fde-data-forge Exact row-key defect scoring with published recall, precision, F1, false positives and misses
rag-eval-bench Retrieval metrics, unanswerable-question scoring, refusal behavior, a versioned judge rubric and a blinded human-agreement workflow
air-gap-deploy-kit Explicit offline bundles, standalone transfer verification and fail-closed rehearsal with image pulls and networking disabled
pgdrift Deterministic Postgres schema snapshots and categorized drift detection for CI and field debugging

🛰️ Field & Edge Operations

Built around the constraint that the network may not be there and the data may not be clean.

Project What it proves
field-sync-notebook Offline-first sync with local queuing and degraded-connectivity recovery
edge-lab-console Full-stack telemetry + config-drift tracking for edge deployments
telemetry-replay-studio Run comparison, outcome scoring, and handoff export for post-mission analysis

🤖 AI Agent Tooling

Project What it proves
agent-security-kit Guardrails for AI coding agents — secret-leak blocking, dangerous-command hooks
ollama-model-router Automatic model selection by prompt type; no manual switching
claude-code-recipes Production patterns for the Claude Code SDK: rate limits, memory, MCP, parallelism
ai-ops-runbook Paste an alert, get a structured operational runbook — LLM-generated, ops-reviewed
workspace-search Local-first RAG over personal docs; no cloud, no account required
awesome-ollama-models Curated free Ollama-compatible cloud models with benchmarks and use cases

⚙️ Automation & Infrastructure Labs

Training environments for automation and infra depth. No real hardware required.

Project What it proves
intent-config-workbench Intent → deterministic config artifacts with reviewable diffs
ansible-convergence-lab Ansible automation training without a control node or real hosts
ansible-wsl-playbook-runner GUI runner for Ansible playbooks on WSL — bridges the Windows sysadmin gap
gitlab-ci-study-lab Hands-on GitLab CI/CD lab with synthetic pipelines
robot-basics / robot-intermediate / robot-advanced Robot Framework training series — 1/3, 2/3, 3/3
cli-device-sim Network device CLI simulator for automation testing without real gear
win-netlab-foundation Windows-first network automation lab — AD, DNS/DHCP, PowerShell baseline
day0-boot-sim DHCP/TFTP provisioning simulator for day-zero device onboarding

🔧 Repo Hygiene Tools

Project What it does
git-clean-history Removes secrets across full git history — AWS, Azure, GCP, OpenAI, Stripe, GitHub
repo-first-aid Audits public-repo trust signals before strangers find the problems
repo-flight-deck Repo health dashboard: CI status, security alerts, PR risk scoring
gh-sentinel GitHub activity monitor and alert CLI

🛠️ Stack

Languages: Python · TypeScript · JavaScript · React · Bash · PowerShell

AI / ML: LLMs · edge inference · Ollama · OpenAI · Claude SDK · RAG · prompt engineering

Cloud / infra: AWS · Docker · Ansible · GitHub Actions · GitLab CI

Networking / sysadmin: Cisco IOS · Active Directory · DNS/DHCP · WSL

Testing & QA: Robot Framework · pytest · automated regression pipelines


🎓 Background

Certifications  ·  CSPO · CSM · AWS Cloud Practitioner · CCNA · CompTIA Security+

  • B.S. Business Administration, Computer Information Systems — Cal State Fresno
  • Software Engineering Certificate (Full-Stack: Python, React, Flask, JS, SQL) — Flatiron School
  • IT Specialist (25B MOS), U.S. Army Cyber Center of Excellence — Distinguished Honor Graduate

snake animation

profile views


Find one of these useful? Star the repo. PRs welcome. Open an issue first for anything bigger than a typo fix.

Pinned Loading

  1. awesome-ollama-models awesome-ollama-models Public

    Curated list of the best free Ollama cloud models with benchmarks and use case recommendations

    16

  2. acme-parts-cloud acme-parts-cloud Public

    Synthetic parts catalog API with intentional data quality defects — seed-driven messiness engine for testing data pipelines

    Python

  3. ai-ops-runbook ai-ops-runbook Public

    Paste an alert or incident — get a structured operational runbook in seconds.

    TypeScript

  4. air-gap-deploy-kit air-gap-deploy-kit Public

    Offline deployment toolkit for the acme-parts-cloud stack — bundles Docker images, Python wheels, and Ollama models for air-gapped installs

    Python

  5. kubedrift kubedrift Public

    Kubernetes cluster state snapshot and drift detector

    Python

  6. recoverops-lab recoverops-lab Public

    Automated PostgreSQL disaster-recovery rehearsal and evidence lab

    Python