Karim Daaboul
AI Enablement • Automation • Developer Experience

I build AI systems that ship.

LLM workflows, internal copilots, and automation with guardrails built for messy ops.

My Role & Focus

I enable teams to adopt AI and automation by building workflows, creating documentation, and training non-technical users. My focus is on making complex technology accessible and practical for real-world business needs.

I specialize in LLM workflows, RAG systems, and automation pipelines that reduce manual work and improve consistency. Whether it's extracting data from unstructured documents or building internal knowledge assistants, I bridge the gap between AI capabilities and business outcomes.

Work Experience

AI Enablement & Automation Lead

First Marble and GraniteRemote

Jun 2025 – Present

  • Built LLM workflows, internal chatbots, and automations for inquiry/quote intake, document parsing, routing, and ops support
  • Implemented guardrails: prompt standards, logging, and human-in-the-loop review
  • Enabled adoption with documentation + training for non-technical teams

Finance Assistant

University of Toronto — Finance Learning Centre

Jan 2023 – Jun 2025

  • Python data cleaning and statistical reporting for headcount datasets
  • Created guides/KB for tools and recurring workflows
  • Advised, guide and support users and faculty for hundreds of software applications available.

Professional Development & Learning Centre Assistant

University of Toronto — Department of Management

Oct 2023 – Jun 2025

  • Organized and led employer engagement workshops in collaboration with EY, Deloitte, and KPMG, personally bringing in industry professionals to share insights on career paths, recruitment expectations, and resume best practices
  • Earned recognition through the Professional Skills Development Program (PSDP), completing over 45 hours of training in professional communication, team leadership, and decision-making under uncertainty.

President — Middle Eastern Students' Association (MESA)

University of Toronto Mississauga

Jan 2023 – Aug 2025

  • Led a multi-functional student organization delivering large-scale cultural and educational events, overseeing strategy, operations, and execution
  • Managed budgeting, financial planning, and post-event performance analysis to improve cost efficiency and funding allocation
  • Built and maintained sponsor and partner relationships, negotiating contracts and coordinating vendors across multiple stakeholders

Skills

AI Enablement

LLM WorkflowsRAG SystemsPrompt EngineeringVector StoresEmbeddingsGuardrails

Automation

n8nWorkflow AutomationAPI IntegrationDocument ProcessingData Pipelines

Data

PythonpandasData CleaningStatistical ReportingExcel/CSV

Developer Experience

Next.jsTypeScriptDocumentationTrainingKnowledge Bases

Tools

GitMicrosoft 365JSON SchemasLogging/MonitoringRole-based Access

Systems Architecture

System DesignDeterministic PipelinesFailure HandlingHuman-in-the-Loop

Systems I've Built

Production AI systems with guardrails and failure handling

AI Inquiry-to-Quote Workflow
First Marble and Granite

AI Inquiry-to-Quote Workflow

Automated email + attachment intake into structured quote-ready data.

Goal

Automate inquiry processing to reduce manual work and standardize quote intake.

Constraints

Must handle unstructured emails and various attachment formats

Guardrails

JSON schema validation for all extracted fields, Confidence thresholds trigger human review

n8nOpenAIMicrosoft 365
View Project →
First Marble and Granite

AI-Driven Workflow Orchestration

Orchestrates multi-step AI and automation workflows by combining deterministic logic with LLM-driven steps—handling branching, failures, and human review across production pipelines.

Goal

Orchestrates multi-step AI and automation workflows by combining deterministic logic with LLM-driven steps—handling branching, failures, and human review across production pipelines.

Constraints

Must combine deterministic rules with AI steps to avoid brittle automation

Guardrails

Conditional logic ensures deterministic branching between AI and rule-based steps, Failure handling prevents partial failures from breaking entire workflows

n8nAPI integrationsLLMs
View Project →
Website-Aware AI Chatbot
First Marble and Granite

Website-Aware AI Chatbot

Provides instant, accurate answers to customer and internal questions by grounding responses in the live website content.

Goal

Provide instant, accurate answers to customer and internal questions by grounding responses in the live website content.

Constraints

Must automatically reflect new or updated site content without manual retraining

Guardrails

Retrieval grounding ensures answers come only from indexed sources, Fallback handling for unavailable information

Next.jsContent ingestionRetrieval grounding
View Project →

Pipeline Example

Interactive demonstration of an AI-powered inquiry-to-quote workflow with guardrails

📚 How to use this demo:

  • Auto Mode: Watch the pipeline run automatically with 4-5 second pauses between steps
  • Step-by-Step Mode: Click through each step manually at your own pace (recommended for learning)
1

Intake

2

Extract

3

Validate

4

Route

5

Output

Let's Work Together

I'm always open to discussing new opportunities and interesting projects.

Get In Touch