Skip to main content
    Oliver Rubio · AI systems, automation and integrations
    Oliver Rubio, AI Automation and Systems Engineer

    I turn operational bottlenecks into production AI systems.

    I design agents, integrations and data workflows end to end. At Listing Profit, that meant 95% faster campaign production, 97% lower cost and a 15-point SLA improvement.

    Download résuméDiscuss a system →
    Bogotá, ColombiaOpen to remote roles in LATAM, US contractor, and global teams
    Operating Equations · Listing Profit

    Measurable systems evidence.

    Production deltas achieved through retryable workflow orchestration, strict output validation, and automated quality gates.

    Turnaround Time
    4h → 5mListing Profit AI creative pipeline
    Unit Cost / Campaign
    ≈$300 → ≈$8Structured n8n + OpenAI asset generation
    On-Time Delivery
    +15 pp SLAIdempotency keys, retries & dead-letter queues
    Ruta Delta · Listing Profit Creative Turnaround
    Before → Outcome
    Selected: Before — Manual copywriting, prompt iterations, and unstructured formatting per campaign.Listing Profit evidence
    Production Case Studies

    Featured systems.

    Three deep-dive production architectures demonstrating multi-agent orchestration, structured data validation, and resilient automation pipelines.

    01/Listing Profit

    AI Marketing Creative Engine

    GTM Engineer | AI Automation Engineer · Aug 2024 – Nov 2025

    Measured deltas:
    Turnaround4h → 5m
    Unit Cost≈$300 → ≈$8
    Delivery SLA+15 pp
    Read system breakdown

    Problem & Bottleneck

    Listing Profit's manual campaign production required copywriting, formatting, and multi-channel asset delivery. Each campaign took about four hours and cost about $300 to produce, creating an operational bottleneck as volume grew.

    Intervention & System

    Built a production n8n and OpenAI pipeline that accepts campaign inputs, generates schema-constrained content, applies QA checkpoints, and prepares multiple delivery formats. The surrounding integration layer connected campaign and lead workflows to Follow Up Boss, GoHighLevel, and webhook endpoints.

    Evidence & Outcome

    Résumé-recorded production outcomes: campaign turnaround fell from 4 hours to 5 minutes, cost fell from about $300 to about $8, and the broader Listing Profit integration work improved operational SLA by 15 percentage points.

    n8nOpenAI APIsStructured OutputsPostgreSQLSupabaseWebhooksFollow Up BossGoHighLevel
    02/Roads Consulting Group US

    marketingAI

    Automation Engineer / Full-Stack AI Lead · Nov 2025 – Jul 2026

    Read system breakdown

    Problem & Bottleneck

    Marketing workflows crossed AI generation, brand context, approvals, publishing, and channel integrations. Fragmented scripts could not provide one auditable path from request to reviewed output and asynchronous delivery.

    Intervention & System

    Built a React and TypeScript web application with a Python FastAPI backend and Pydantic AI agents. The current production topology runs web and API services in Cloudflare Containers, dispatches asynchronous work through Cloudflare Queues, and persists data and media in Neon PostgreSQL and R2.

    Evidence & Outcome

    Shipped and maintained in live production at marketingai.roadscg.com, supporting multi-agent marketing generation workflows with strict schema safety and low-latency edge delivery.

    ReactTypeScriptPythonFastAPIPydantic AIDockerCloudflare ContainersCloudflare QueuesNeon PostgreSQLR2
    03/Open-Source AI Tooling

    Agent Bridge

    Author & Systems Engineer · 2026

    Read system breakdown

    Problem & Bottleneck

    Terminal-based AI agents and CLI copilots typically operate in isolated session silos. Passing state, sub-task delegations, and execution context between disparate agent runtimes required cumbersome manual copying or fragile temporary files.

    Intervention & System

    Designed a Python CLI and protocol with SQLite-backed messages and tasks, Unix-domain socket inbox delivery as the primary transport, and explicit tmux prompt injection as a fallback. Runs, policies, delivery states, acknowledgements, and task claims remain durable locally.

    Evidence & Outcome

    Published as an open-source public repository on GitHub (github.com/oliveralerubio/agent-bridge) with complete test suites and documentation.

    PythonSQLiteUnix Domain SocketsLocal-First ArchitectureIPC / Session HandoffMIT License
    Additional Engineering Work

    Selected builds & tools.

    Go Aria AI

    Lead generation & multi-channel outreach engine with Retell AI voice agents, n8n, Supabase, and GoHighLevel CRM synchronization.

    Visit GoAria.ai

    FastClaw

    Shipped product

    Shipped modern React and TypeScript SaaS product with accessible UI primitives, fast client-side performance, and modular architecture.

    Visit FastClaw

    Clay Waterfall Enrichment Pipeline

    Multi-vendor waterfall data enrichment combining Apollo, Apify, and LinkedIn Sales Navigator for optimal ICP coverage and CRM sync.

    Engineering Architecture

    Systems architecture & lifecycle.

    How I architect production AI pipelines from event ingestion to observability, ensuring high reliability and deterministic execution.

    Architectural lifecycle diagram representing Trigger, State, Agent, Human gate, Delivery, and Observability stages.
    Stage 01 · Trigger

    Idempotent Webhooks & Queues

    Incoming API or CRM events are authenticated, deduplicated with idempotency keys, and queued before downstream work begins.

    Operational Safeguards & Mechanisms:
    • Webhook authentication
    • Idempotency keys
    • Rate-limited ingestion queue

    Full Systems Stage Lifecycle

    1. Trigger (Event Ingestion): Idempotent Webhooks & Queues. Incoming API or CRM events are authenticated, deduplicated with idempotency keys, and queued before downstream work begins.
    2. State (Persistence & Context): PostgreSQL & Row-Level Security. Workflow state, approval history, and execution records live in PostgreSQL with account-scoped access and auditable status changes.
    3. Agent (Reasoning & Schemas): Pydantic AI & Structured Outputs. Agents operate through typed schemas, explicit tools, and provider fallbacks so downstream steps receive validated data.
    4. Human gate (Review & Approval): Heuristic Flags & Checkpoints. High-impact actions, uncertain outputs, and approval-required assets stop for human review before delivery.
    5. Delivery (Multi-Channel Dispatch): API Sync & Delivery Cadences. Approved outputs sync to CRM, messaging, storage, or publishing APIs with exponential backoff and dead-letter handling.
    6. Observability (Telemetry & Tracing): Logfire, Tracing & Alerting. Critical steps emit traces and structured events for latency, cost, failures, and SLA review.
    Career Chronology

    Work experience.

    Track record delivering production automation, autonomous agent integrations, SaaS platforms, and enterprise data pipelines for US and LATAM teams.

    1. 01 · Nov 2025 – Jul 2026

      Automation Engineer

      Roads Consulting Group US
      Remote (US)
      Outcome:Shipped marketingAI into live production and stabilized core cross-platform event integrations.
      Scope & Ownership

      Marketing, finance, and operational automation across multi-agent AI and enterprise APIs.

      Key Technical Deliverables
      • Built marketingai.roadscg.com, a production full-stack AI SaaS using React, Python, Cloudflare, Docker, and Pydantic AI.
      • Engineered automated workflows across marketing, finance, and recruiting using Python, n8n, Azure, and REST APIs.
      • Integrated Meta Developers Platform and core business systems with webhook pipelines and human-in-the-loop validation.
      • Hardened failure modes, authored technical runbooks, and translated complex client requirements into maintainable architectures.
      Pythonn8nPydantic AIFastAPIReactCloudflareDockerMeta Developers PlatformREST APIs
    2. 02 · Aug 2024 – Nov 2025

      GTM Engineer | AI Automation Engineer

      Listing Profit
      Remote (US-based Real Estate SaaS)
      Outcome:95% faster campaign production, 97% lower unit cost, and +15 pp operational delivery SLA.
      Scope & Ownership

      AI outreach agents, creative pipeline automation, data engineering, and operational SLA hardening.

      Key Technical Deliverables
      • Engineered AI creative and outreach pipelines in Python using FastAPI, Pydantic AI, LangGraph, and PostgreSQL.
      • Integrated Follow Up Boss, GoHighLevel, Retell AI, OpenAI, Stripe, Supabase/Postgres, and webhook endpoints.
      • Reduced campaign production time by 95% (4h → 5m) and cost by 97% (~$300 → ~$8); improved delivery SLA by 15 percentage points.
      • Implemented idempotency keys, exponential retries, dead-letter queues, and multi-tenant PostgreSQL with Row-Level Security.
      • Architected Postgres-to-Snowflake ELT pipelines and dbt models with data-quality tests for reporting dashboards.
      PythonFastAPIPydantic AILangGraphPostgreSQLSupabase / RLSSnowflakedbtRetell AIn8n
    3. 03 · Aug 2023 – Dec 2024

      Data Analyst / Analytics Engineer

      TaskUs
      Remote (Bogotá)
      Outcome:Standardized metric definitions and automated recurring operational reporting flows.
      Scope & Ownership

      Operational analytics, data validation pipelines, and executive BI reporting.

      Key Technical Deliverables
      • Analyzed operational and people-analytics datasets using SQL, Python (pandas), Excel, and Power BI.
      • Designed scheduled data pipelines, automated validation checks, and formalized business metric definitions.
      • Delivered KPI dashboards and operational reporting for executive stakeholders.
      SQLPythonpandasExcelPower BIData ModelingValidation Checks
    Technical Capabilities & Scope

    Engineering deliverables.

    Specialized in multi-agent orchestration, event-driven integrations, robust database architecture, and full-stack AI execution.

    Capability 01

    Multi-Agent Orchestration & Systems

    Designing autonomous agent networks, structured LLM pipelines, and stateful multi-agent workflows with schema-safe tool calling.

    Core Deliverables:
    • Pydantic AI & LangGraph agent state machines
    • Strict JSON schema output validation
    • Multi-model routing & token cost controls
    • Context compaction & long-running agent state
    PythonFastAPIPydantic AILangGraphLangChainOpenAI APIsRetell AI
    Capability 02

    Event-Driven Automation & Queues

    Connecting disparate SaaS and custom backends with resilient webhooks, message queues, and automated error handling.

    Core Deliverables:
    • n8n & Make workflow pipelines with dead-letter routing
    • Idempotency keys and exponential backoff retries
    • Human-in-the-loop approval gates for high-risk actions
    • CRM & payments integrations (GoHighLevel, Stripe, Meta)
    n8nMake.comREST APIsGraphQLWebhooksStripeGoHighLevelFollow Up Boss
    Capability 03

    Database Architecture & Data Pipelines

    Building reliable relational schemas, multi-tenant databases with Row-Level Security, and observable analytical ELT pipelines.

    Core Deliverables:
    • Multi-tenant PostgreSQL schemas with Supabase RLS
    • ELT pipelines from Postgres to Snowflake via dbt
    • Data quality assertions and automated schema tests
    • Redis caching and background task scheduling
    SQLPostgreSQLSupabase / RLSSnowflakedbtRedisPython / pandas
    Capability 04

    Full-Stack AI SaaS & Edge Routing

    Shipping accessible React applications, containerized backend services, and low-latency edge network configurations.

    Core Deliverables:
    • Accessible React + TypeScript web applications
    • Docker containerization & zero-downtime deployments
    • Cloudflare edge proxying, caching, and SSL termination
    • Real-time tracing with Logfire & OpenTelemetry
    ReactTypeScriptTailwind CSSDockerCloudflareFastAPIAzureVite
    Origins & Foundation

    Sociology to systems.

    How formal training in quantitative research and organizational modeling translates into resilient AI architectures.

    “I learned to model human systems before automating software systems.”

    I learned to model human systems before automating software systems. My academic background in quantitative sociology gave me a rigorous framework for understanding organizational bottlenecks, state transitions, and incentive structures—the exact foundations needed to design reliable AI automation architectures.

    Education

    M.A. in Sociology

    Universidad de los Andes

    Bogotá, Colombia · 2019 – 2021

    GPA 4.5 / 5.0 · Full Beca Impacto País Scholarship recipient

    Honors & Scholarships
    • Beca Impacto País ScholarshipUniversidad de los Andes
    • Faculty Excellence ScholarshipUniversidad de los Andes
    • ICFES Saber Pro Top Performers CohortNational Examination Board (Colombia)
    • ICFES Saber 11 Top ScoreSchool-Level Ranking
    Languages & Setup
    SpanishNative / Bilingual
    EnglishProfessional Working Proficiency

    Bogotá, Colombia (UTC-5). Open to US contractor and global remote engineering roles. Async-first collaboration.

    Direct Inquiries & Architecture

    Bring me the bottleneck.

    Whether you need to scale autonomous multi-agent pipelines, harden brittle integrations, or turn manual real-estate or GTM workflows into reliable production software, let's talk directly.

    Email Oliver (oc25ale1917@gmail.com)
    Availability & Direct Links
    StatusOpen to Remote / US Contractor
    LocationBogotá, Colombia (UTC-5)
    GitHubgithub.com/oliverr
    LinkedInlinkedin.com/in/oliver-rubio
    Résumé PDFDownload PDF