larasaints.github.io 个人作品集：AI 自动化与全栈项目展示

Digital Smith | AI Automations Specialist | GHL, N8N, Zapier, Make, API & Webhook, GHL's Conversation AI Integration, RAG, LLMs, Error Handling & Escalation Workflow

I transform concepts into high-performing digital realities. By blending intelligent AI workflows, strategic graphic design, and full-stack software development, I build end-to-end solutions that scale. Let’s partner up to streamline your operations and drive your business forward.

📬 Let's Connect:

• 📧 Email: lrjoysantos@gmail.com
• 💼 LinkedIn: linkedin.com/in/larajoysantos
• 🟢 Upwork Profile: Hire Me on Upwork
• 🇵🇭 Onlinejobs.ph: View Profile Credentials

🛠️ Tech Stack & Capabilities

• 🧠 AI & Large Language Models: OpenAI API • Claude • Gemini AI • DeepSeek • Grok
• 🔌 Enterprise Workflow Automation: Make.com (Integromat) • n8n • Zapier
• 💻 Low-Code & Full-Stack Prototyping: Lovable.io • Supabase Database Integration • Web App Architecture • Rapid Prototyping
• 🎨 System Design & Asset Creation: UI Component Design Systems • Template Engineering • Brand Visual Identity • Digital Asset Architecture
• 🕹️ Logic & Interactive Media: Event-Driven Programming • Logic Architecture • Game Mechanics Design • Interactive Storytelling • Motion Graphics & Video Production
• ⚙️ Backend Dev & Version Control: GitHub • REST API Integration

📂 FEATURED PROJECTS

Automations

I design interconnected automation systems that combine AI, CRM platforms, structured data, workflow logic, exception handling, and operational dashboards. My projects focus on turning manual business processes into repeatable, trackable systems with clear escalation paths and measurable workflow states.

• 💻 View All My Automation Projects & Workflow Breakdown

🏢 UrbanBite: Multi-Zap Franchise Operations & Compliance Automation Engine

• The Problem: Managing franchise operations across multiple locations manually leads to operational bottlenecks. Inspection results can be difficult to track, critical business permits can approach expiration unnoticed, and missing monthly reports require compliance teams to manually cross-reference records and follow up with operators.

• The Solution: A centralized, multi-Zap automated operations architecture that manages the franchise compliance lifecycle. The system captures form submissions, updates tracking databases, monitors deadlines using scheduled workflows, triggers operational escalations, and feeds status metrics into a centralized operations dashboard.

• Ecosystem Architecture — Five Automation Workflows:

  • ZAP 01 — New Compliance Submission: Captures submissions, creates compliance records, identifies reported issues, and triggers appropriate alerts.
  • ZAP 02 — Issue Resolution & Corrective Action: Processes resolution submissions, updates issue records, and triggers follow-up notifications when required.
  • ZAP 03 — Overdue Issue Monitoring & Escalation: Identifies overdue issues and sends a consolidated escalation report.
  • ZAP 04 — Document Expiration Monitoring: Monitors franchise documents and identifies expired or upcoming expirations requiring renewal action.
  • ZAP 05 — Missing Monthly Compliance Submission: Checks active franchises for missing monthly submissions and sends consolidated reminders while preventing duplicate alerts.

• Key Engineering Metrics: Designed a 5-Zap interconnected ecosystem, engineered a unified Google Sheets relational data layer, integrated automated deadline monitoring, and built multi-condition operational escalation logic.

• Code & Architecture:

  • 💻 View UB repository, diagram and system flows
  • 🎥Watch the Loom Workflow Demonstration

🎧 AI Customer Support Automation & Intelligent Ticket Triage Engine

• The Problem: Customer support teams often spend significant time manually reading incoming requests, categorizing issues, creating tickets, drafting responses, and identifying cases that require escalation.

• The Solution: Built an AI-assisted support intake and triage workflow using n8n and Google Gemini. The system receives customer messages through a webhook, checks for duplicates, classifies the request, generates a draft response, logs structured ticket data in Airtable, and routes sensitive or uncertain requests for human review.

• AI & Automation Logic:

  • Classifies requests into Billing, Access, Scheduling, Technical Support, General Question, or Refund Request.
  • Generates a confidence score and classification reason.
  • Creates an AI-generated customer response draft.
  • Automatically escalates refund requests and low-confidence cases.
  • Sends internal Gmail notifications when human review is required.
  • Tracks ticket status and processing timestamps in Airtable.
  • Uses a separate n8n error workflow for error logging and internal alerts.
  • Prevents duplicate processing through Message ID validation.

• Key Engineering Metrics: Event-driven webhook architecture, AI-powered classification, deterministic human-review rules, structured Airtable ticket logging, duplicate prevention, and automated error handling.

• Code & Architecture:

  • 💻 View Customer Support repository, workflow documentation and testing
  • 🎥Watch the Loom Workflow Demonstration

🤖 Hands-Free AI-Powered Facebook Leads Automation Engine

• The Problem: Real estate lead intake introduces major response delays when done manually. Hot buyer interest can cool down within minutes, while manual follow-ups can create calendar conflicts and inconsistent customer communication.

• The Solution: Built a real-time, event-driven pipeline on Make.com connected to Gemini AI. The engine captures Facebook Lead Ads data packets, sorts client profiles by property choice using filters, logs records to Google Sheets, checks calendar availability, and drafts custom property proposals sent to the buyer via Gmail.

• Key Engineering Metrics: Sub-2-minute automated lead response times, real-time webhook data capture, parallel multi-property path routing, and automatic error-handling retry guards.

• Code & Architecture:

  • 💻 View Facebook Leads Documentation

🏓 THE KITCHEN DECK : Mobile-Responsive Web Application, Production-Grade Product

• The Problem: Traditional sports training regimens suffer from low compliance due to two primary friction points:

  1. Cognitive Load on Court: Players must remember complex drill layouts, scoring rules, and target zones while physically exerting themselves.
  2. Lack of Feedback Loops: Static drill sheets offer no immediate psychological reward, leading to rapid dropout rates and unstructured practice sessions.

• The Solution: A high-contrast, touch-optimized visual engine that translates structural sports drills into gamified digital cards.

  1. Visual Frameworks: Breaks down spatial movement patterns into glanceable, high-fidelity card layouts.
  2. Behavioral Mechanics: Replaces mental tracking with real-time feedback loops, XP counters, and level progression models to increase user engagement.

• Core & Architecture: The underlying source code, database structures, custom visual card templates, and corporate branding for this project are proprietary and hosted in a private repository to protect intellectual property. Permission is granted strictly for public portfolio demonstration, recruitment evaluation, and educational review via the live published site. No part of the visual assets or architecture may be replicated*