Portfolio - Pedro Alejandro Bestard Hernandez

About Me

Hello! I’m Pedro Bestard, a Data Scientist with a strong foundation in Natural Language Processing (NLP), Machine Learning, and statistical modeling. I specialize in turning structured and unstructured data into actionable insights that improve decision-making.

⚙️ Tools & Technologies:
🟦 I use Python to automate end-to-end data pipelines and model development.

🟦 I’m skilled in SQL for retrieving data.

🟦 I also work with tools like Hugging Face, LangChain, Streamlit and Gradio to build and deploy AI applications.

📌 Areas of Expertise:
🟦 Scaling NLP solutions and automating manual analysis workflows.

🟦 Developing predictive models for churn, satisfaction, and text-based classification.

🟦 Designing dashboards and feedback systems for real-time business insight.

I’m passionate about building systems that turn data into impact, and I love sharing what I learn. If you're curious about how analytics can drive real results, feel free to explore my work or connect with me!

Key Skills

Machine Learning

Deep Learning

LLMs

GenAI

A/B Testing

Computer Vision

Natural Language Processing

My Projects

Expense-Tracking-System

Domain/Function: Personal Finance / Financial Technology (FinTech)

Health Insurance Cost Predictor

Domain/Function: Health Insurance

Lauki Finance: Credit Risk Modeling

Domain/Function: Banking/Credit Risk

My Experience

• Redesigned and scaled Telcel’s NLP program from 2 to 15+ projects, enabling 360° customer insight across all channels.

• Increased classification coverage by 30% and boosted accuracy by 20% through improved text cleaning weighting, and business logic.

• Raised company-wide NPS by ∼30% by applying text analytics and causal impact modeling to identify key service drivers.

• Improved customer satisfaction scores by 30% using machine learning and survey optimization to uncover root causes of dissatisfaction.

• Cut analysis delivery time from 2 weeks to 1 day by automating the pipeline using Python, APIs, and SAS.

• Fully automated daily manual tasks, from data gathering and cleaning to survey distribution and dashboard generation.

• Built a national analytics dashboard integrating market share, network quality, and socioeconomic data at ZIP-code level. Enabled 500+ stakeholders to make data-driven decisions.

• Designed a reusable ETL workflow adopted across the analytics team, accelerating data preparation and delivery.

• Created modular SAS pipelines, establishing a standardized process for recurring customer analysis.

• Built 5+ real-time feedback dashboards in Qualtrics, paired with user documentation and training for directors and managers.

• Mentored all new NLP team members, delivering onboarding sessions and team-wide workshops on text analytics and modeling.

• Collaborated cross-functionally with marketing, operations, and governance teams to build the Network Quality project—linking geolocated feedback with network KPIs to detect hidden service issues.

• Initiated a churn prediction model using 100+ features from behavior logs, internal systems, and customer feedback.

Awards & Certificate