Data & AI Professional

Bogotá, Colombia

Juan Felipe García.

Building end-to-end data and AI solutions, from raw data to better decisions.

Computer Science & Artificial Intelligence student working across data science, analytics, cloud data engineering and applied AI.

I connect the whole path: collecting and processing data, analyzing it, building models and AI applications, deploying and automating them, and tying the result to a business decision.

View Projects Get in Touch Download CV

01
DataPipelines and processing
02
AnalyticsUnderstand the problem
03
Models & AIBuild the solution
04
DecisionsDeploy, automate, measure

Selected work

From prediction to a business decision.

Data Science & ML · Banking

Retention Copilot

Ranking customers by churn probability ignores how much value a call could preserve, and a fixed 0.5 threshold doesn’t respond to changes in campaign cost or effectiveness. The question is which customers to contact to maximize the expected benefit of a retention campaign.

Python
pandas
NumPy
Scikit-learn
XGBoost
Streamlit
GitHub Actions

Read case study GitHub

Retention Copilot Streamlit app (Spanish interface): campaign parameters in the sidebar, a simulated campaign recommendation with customers to contact, expected preserved revenue, campaign cost and expected incremental benefit, and the start of a call list prioritized by churn probability. — Interactive scenario simulator. Dashboard figures change with campaign assumptions and are simulated; the evaluated holdout results are reported separately below.

Model evaluation & campaign simulation

customers analyzed: 10,000
ROC-AUC (held-out test): 0.855
Average Precision (held-out test): 0.698
fewer contacts than contacting everyone: 31.5%
more simulated incremental profit than contacting everyone: US$6,405

Financial results are simulations under explicit assumptions (US$15 per contact, 30% retention success). They are not observed bank revenue or causal effects, and annual customer value is not full CLV.

Professional experience

Real operations. Real impact.

Putting data to work in telecommunications infrastructure.

Telecommunications / Operations

Azteca Comunicaciones

NOC Analyst Intern

Jan 2026 – Jul 2026 · Bogotá, Colombia

Automated an equipment inventory workflow, reducing inventory management time by approximately 35%.
Built an application for management, monitoring and statistical analysis supporting network operations across more than 900 municipalities.
Supported daily Network Operations Center (NOC) activities, including monitoring and management of telecommunications infrastructure.

reduction in inventory management time: ≈35%
municipalities supported: 900+

Operations Analytics

Management, monitoring and statistical analysis

More projects

More ways to put data to work.

Analytics, data engineering and applied AI.

Analytics & BI · Retail

Vendor Performance Analysis

A vendor-brand data model in SQLite, statistical analysis in Python and a Power BI dashboard for purchasing, sales and inventory.

Python
pandas
SQLite
SQL
Jupyter
Matplotlib
Seaborn
SciPy
pytest
Power BI

GitHub

Cloud & Data Engineering · Markets

Stock ETL Pipeline

A scheduled ETL from Yahoo Finance into BigQuery for 31 Colombian and U.S. tickers, served through FastAPI, Streamlit and a Claude agent.

Python
pandas
yfinance
BigQuery
FastAPI
Docker
Cloud Run
GitHub Actions
Streamlit
Plotly
Anthropic Claude API

GitHub

MediGuide Chat interface from the medical RAG prototype, showing an education-only disclaimer and a sample conversation answering general questions about SARS and VSV.

Applied AI & GenAI · RAG

Portfolio RAG prototype

Medical Chatbot with LLMs

A portfolio RAG prototype that answers general medical questions from a reference book, built with LangChain, Pinecone and GPT-4o.

Python
LangChain
HuggingFace
Pinecone
GPT-4o
Flask
Docker
AWS ECR
AWS EC2
GitHub Actions

GitHub

Power BI Bank Loan Report, Summary page: KPI cards for total loan applications, funded amount, amount received, average interest rate and average DTI with MTD and MoM values; good and bad loan breakdowns; and a loan-status table.

Analytics & BI · Banking

Bank Loan Analysis

A three-page Power BI report on 38,576 loans from 2021, validated against SQL, from portfolio KPIs down to individual records.

Power BI
DAX
SQL

GitHub

About me

Start with a problem.
Build toward an outcome.

Based in Bogotá, Colombia

I’m Juan Felipe, a Computer Science and Artificial Intelligence student focused on data and AI.

I work across the whole chain: getting and cleaning data, analyzing it, training models or building LLM applications, deploying them to the cloud, and connecting the result to a decision someone has to make.

From fiber infrastructure operations to banking analytics, I care about what a solution makes possible and how its impact is measured.

Technical toolkit

The tools behind the work.

Four connected areas, one end-to-end workflow.

Data Science & Machine Learning

Python
pandas
NumPy
Scikit-learn
XGBoost

Analytics & Business Intelligence

SQL
Power BI
Excel

Cloud & Data Engineering

BigQuery
Cloud Run
FastAPI
REST APIs
Docker
GitHub Actions
AWS

Applied AI & Generative AI

LLMs
RAG
LangChain
Pinecone
Prompt Engineering
n8n

Let’s connect

Let’s build something
with data & AI.

Have a problem worth solving? Let’s talk.

GitHub LinkedIn Email Download CV

Juan Felipe García.

Building end-to-end data and AI solutions, from raw data to better decisions.

From prediction to a business decision.

Retention Copilot

Real operations. Real impact.

Azteca Comunicaciones

Operations Analytics

More ways to put data to work.

Vendor Performance Analysis

Stock ETL Pipeline

Medical Chatbot with LLMs

Bank Loan Analysis

Start with a problem.Build toward an outcome.

The tools behind the work.

Data Science & Machine Learning

Analytics & Business Intelligence

Cloud & Data Engineering

Applied AI & Generative AI

Let’s build somethingwith data & AI.

Start with a problem.
Build toward an outcome.

Let’s build something
with data & AI.