Underground, something is waiting.
Ariel Rubinstein
Independent AI Automation Consultant
I build AI automation systems — AI agents write the code; I architect, direct, verify, and operate — with a 4-year production track record automating finance operations for an air-cargo GSA.
Scroll to grow
First fruit.
Featured builds
Systems I built that are running in production right now — finance back-offices, knowledge infrastructure, and the automation that keeps my own operation honest.
BSP reconciliation platform
Production finance system · air-cargo GSA
Automated the monthly IATA BSP records cycle — download, extraction, reporting — replacing a ~4–5-hour manual monthly process. Now consolidating the surrounding legacy stack (15–20 macro-driven Excel files, 9 years of records across 3 companies, 3 banks, and 4 currencies) into a single Next.js + PostgreSQL platform, built with AI agents under my direction (in progress).
- ~4–5 hours of manual work per month removed
- 9 years of records, 4 currencies, one platform
Airline invoice reconciler
Production finance system · air-cargo GSA
Python ZIP-to-PDF extraction feeding a Power Query / Power Pivot model that compares cargo billing exports against airline invoices and surfaces the discrepancies a human needs to decide on. Heroku-hosted; outputs analyst-ready decision-point reports instead of raw spreadsheets.
- ~4 years in continuous production
- Eyeballing spreadsheets → decision-point reports
Job-hunt digest pipeline
Personal automation · nightly production
Nightly scans across job boards and ATS APIs (Greenhouse, Ashby, Lever) covering 50–200 companies per run, LLM fit-scoring against a locked personal rubric, and a ranked morning digest. Three candidate profiles run in production with a FastAPI health dashboard.
- Runs every night, unattended
- Deterministic pre-filters + LLM scoring rubric
Document-intelligence stack
Knowledge infrastructure · self-hosted
Kreuzberg text extraction feeding a LightRAG knowledge graph with semantic search over 50+ documents, on a nightly ingest. An MCP server exposes four knowledge-base tools to Claude Code agents, so every automation runs grounded in accumulated context instead of from scratch.
- Knowledge graph + hybrid semantic search
- Agents query it as a first-class tool
Telegram agent fleet
Personal automation · self-hosted
A fleet of Telegram bots backed by scheduled agent jobs on a home server: system monitoring, task and calendar management, daily accountability check-ins, and a content-approval gate where a human approves before anything ships.
- Six bots, one hybrid architecture
- Human-in-the-loop by design
Agency operations board
Internal tooling · Next.js
A kanban work surface and cross-client pipeline dashboard for a consulting practice — markdown files stay the canonical store, a web board serves as the management surface, with an API-guarded card system, stalled-work alerts, and cron-driven reporting.
- Markdown-canonical, API-managed
- One board across every client
Leaves unfurl.
How I work
My deepest expertise is the orchestration layer: Claude Code and custom Agent Skills, MCP, n8n, and the SaaS API graph — composed into internal systems that run on a schedule and move the human up to deciding on exceptions.
- AI & agentic engineering — expert
- Claude Code & custom Agent Skills (80+ authored across 17 plugins) · Model Context Protocol (MCP) · n8n · multi-agent orchestration · RAG & knowledge graphs · document intelligence · prompt/context engineering · local LLM hosting
- Hands-on data & automation
- Excel · Power Query / Power Pivot · VBA — four years of daily production use
- AI-directed engineering
- Python · SQL · TypeScript / Next.js · PostgreSQL · Supabase — agents write the code; I architect, review, debug, design schemas, test, deploy, and operate
The common thread: systems that run unattended in production, with a human deciding only on exceptions.
Count the rings.
Experience
Open Sky Cargo GSA
Internal Operations / Automation Engineer · 2022 — Present
Built and operate the reconciliation, invoicing, and finance-modernization stack — production systems on Python, Power Query / Power Pivot, and VBA that I own at system level. Currently leading the migration of the legacy stack to a Next.js + PostgreSQL platform, built with AI agents under my direction (in progress).
MyNewAgent.AI
Independent AI Automation Consultant · May 2025 — Present
Independent AI-automation practice: agentic workflows composing n8n, Claude Code, and MCP; document intelligence and multilingual RAG; local and cloud LLM infrastructure on self-managed servers and Vercel.
RecruitmentFX.com
Talent Acquisition Specialist · 2021 — 2022
Full-cycle technical talent acquisition and candidate pipeline management.
Waserstein Nunez & Foodman (WNF Law, P.L.)
Project Manager · 2017 — 2018
Managed operational projects and KPI tracking for a Florida law firm.
Earlier
Legal, insurance, real-estate & operations roles · 2008 — 2021
Legal, insurance, and real-estate roles in Florida (2008–2016); customer-operations and recruiting roles in Israel following relocation (2018–2021).
BS, Resource Economics — University of Florida · Data Analysis (SQL, Python, Excel/VBA) — Elevation Academy · Google Data Analytics — Coursera
Full bloom.
Get in touch
That's the journey so far — and the part I like best is that every system above is still running. If you're hiring for internal tools, automation, workflow, or business-systems work, I'd like to hear about it.
I also take on select automation projects → mynewagent.ai