Angel Lopez

AI Engineer (GenAI) · Agentic Systems · RAG (pgvector/Chroma)
Mexico · agdaniel019@gmail.com

AI Engineer (GenAI-focused) with a Data Engineering background. Built production agentic systems with tool/function-calling and RAG (OpenAI/Gemini embeddings, Postgres/pgvector, Chroma) shipped as APIs with Python, Django, Docker, and Pydantic. Delivered automation reducing operational workload by up to 60% and supported monetization via usage analytics and pricing redesign. AWS (EC2, S3, IAM). English B2.

Experience

AI Developer

WispHub
GenAI Agentic Systems RAG Production
January 2025 – Present

Built AI functionalities inside the CRM for support and sales teams. Developed agents that execute tools (tool/function-calling) and use RAG to respond with context from the internal knowledge base, reducing response times and improving support quality at scale.

~100
companies adopted self-serve AI features
70%
reduction in first-response time
80%
users reported process improvement (surveys)
~$5
/ 2,000 msgs — cost model for pricing
What I Built
  • Agentic workflows with PydanticAI + tool/function-calling
  • RAG pipeline with internal knowledge base (pgvector + PostgreSQL)
  • Self-serve flow builder to configure agents without code
  • Sales agents to automate first response and lead prioritization
How I Did It
  • Python + Django / FastAPI + Pydantic
  • OpenAI / Gemini (embeddings and completions)
  • Postgres/pgvector as vector store
  • Linux (DigitalOcean) + Supervisor for deployment
  • Structured output validation
AI agent workflow architecture diagram
Agent architecture: message → RAG → decision → response / workflow
Live AI agent interaction via WhatsApp
Agent responding to customer via WhatsApp
Agent configuration interface in CRMinbox
Self-serve agent configuration in CRMinbox

How we measured impact

Post-interaction satisfaction surveys, first-response times logged in the CRM, and flow adoption rate per company (internal platform data).

Cost considerations

~$5 USD per 2,000 messages estimated considering embedding + completion tokens (OpenAI/Gemini). Enables defining customer tiers and pricing decisions.

Deployment

Infrastructure on DigitalOcean (Linux), process management with Supervisor to keep workers active, Docker to isolate services and enable zero-downtime updates.

Screenshots shown with anonymized or illustrative data.

Database Assistant (Intern)

Municipal Secretariat of Public Security and Transit
  • Implemented ETL processes to clean and update the licensing database using DynamoDB in AWS
  • Automated data cleaning and report generation tasks

  • DynamoDB

  • AWS

December 2023 - January 2024

Marketing Developer (Intern)

Zavia ERP
  • Designed and fully developed application to search for potential client information
  • Enabled retrieving client data for sales in seconds

  • Python

  • Customtkinter

  • Pandas

  • BeautifulSoup

September 2023 - November 2023

Marketing Developer (Intern)

Zavia ERP
  • Developed application using web scraping techniques to extract hotel information
  • Facilitated data storage and analysis in CSV and Excel formats

  • Python

  • Requests

  • Pandas

  • BeautifulSoup

  • Numpy

  • Jellyfish

  • Openxlsx

June 2023 - August 2023

Education

Universidad del Caribe

Bachelor's Degree in Engineering
Data Engineering and Organizational Intelligence

Final Grade: 9.23/10

Top GPA in the program

August 2020 - January 2025

Colegio Boston Tikal

High School

Final Grade: 9/10

September 2018 - August 2020

Projects

Plankton Classifier

  • Reduced runtime from 5 hours to 6 seconds; National Winner ANUIES 2025
Python TensorFlow InceptionV3 Anvil
GitHub Demo

WebScraperHotels

  • Reduced data collection from 2 hours to less than 5 minutes
Python BeautifulSoup Pandas
GitHub

Plankton Image Dataset Creation

Complementary project to the Plankton Classifier. Collected ~600,000 images in 40 minutes from IFCB Dashboard and PlanktonNet, organized by class to enrich the original WHOI dataset (2006-2014).

GitHub

Skills

AI/GenAI

OpenAI Gemini LangChain LangGraph PydanticAI RAG pgvector Chroma

Backend

Python Django FastAPI REST Docker

Data

Postgres MySQL DynamoDB ETL/ELT

Infra

Linux Hetzner Supervisor AWS (EC2/S3/IAM)
Programming Languages & Tools

  • Python

  • TensorFlow

  • FastAPI

  • Docker

  • Git

  • Linux

  • AWS

  • MySQL

  • Pandas

  • Github

  • Scikit-learn

  • Numpy

Workflow
  • ETL/ELT Pipeline Design and Implementation
  • Collaboration with Cross-Functional Teams
  • Data Cleaning and Preprocessing
  • Data Visualization
  • Performance Optimization in Big Data Environments

Certifications and Courses

Amazon Educate Badges