Automated Product Tagging
ResNet18 model learned reliable features from limited, imbalanced data and reduced manual catalog tagging effort by ~80%, enabling faster product onboarding for Artiszën Crafts.
About
Neysha Pagán is a strategic Data & AI leader with 14+ years of experience designing and modernizing enterprise-scale data platforms, leading cloud transformation initiatives, and enabling advanced analytics across complex organizations. She specializes in architecting governed Lakehouse ecosystems, aligning data strategy with business objectives, and embedding AI-driven automation to improve decision-making and operational performance.
She holds an M.S. in Applied Data Science from Syracuse University and is currently completing the Chief Data & AI Officer (CDAIO) program at Carnegie Mellon University. Her focus centers on enterprise data governance, AI strategy, and building sustainable, value-driven data organizations that support long-term business growth.
Education
CDAIO Program
- Chief Data & AI Officer Program - Carnegie Mellon University, Heinz College
Graduate
- M.S., Applied Data Science — Syracuse University
Undergraduate
- B.S., Computer Information Systems — University of Puerto Rico at Mayagüez
Skills
Certifications
Microsoft Professional Certified
- Azure Database Administrator Associate — Aug 2022
- Azure Data Fundamentals — Apr 2022
- Azure Fundamentals — Nov 2020
- Administering a SQL Database Infrastructure — Sep 2018
- MTA: Database Fundamentals — Mar 2017
Profisee
- Master Data Architect — Aug 2022
- Master Data System Administrator — Jul 2022
Others
- SAFe 5.0 Agilist — Oct 2022
- IBM Big Data Technologies — 2019
- Microsoft Professional Program: Big Data / Data Science / AI — 2017
Portfolio Summary
Across the five projects, Neysha strengthened end-to-end skills in data engineering, machine learning, and deep learning—while emphasizing governance, automation, and measurable outcomes.
- Data quality first: schema alignment, missing-data handling, and class imbalance drive downstream results.
- Beyond accuracy: solutions must be explainable, fair, reliable, and deployable.
- Communicate impact: clear visuals and storytelling make insights actionable.
- Context matters: domain knowledge + collaboration amplify technical work.
Projects
Click any card to view the full project.
Women in STEM
Scripting for Data Analysis
Analysis of gender disparities in STEM employment and pay equity across U.S. states. Click the card to view the full project.
Product Image Classification (Deep Learning)
Deep Learning in Practice
CNN-based classification of Artiszën Crafts product images (~84.7% accuracy). Click the card to view the full project.
Predictive Maintenance for Industrial Equipment
Big Data Analytics
Machine-failure prediction and anomaly detection with SMOTE and Random Forest. Click the card to view the full project.
Sentiment Classification of Movie Reviews using NLP Techniques
Natural Language Processing
Hybrid TF-IDF + lexicon features for 5-class sentiment on Rotten Tomatoes phrases. Click the card to view the full project.
Childcare Costs vs. Family Income and Employment Data
Advanced Big Data Management
Three-part project integrating childcare prices, labor statistics, and state incomes with Spark, Cassandra/MongoDB, and Kibana dashboards. Click the card to view the full project.
Project Outcomes
Unified Childcare Affordability Pipeline
Built an end-to-end pipeline joining childcare prices (6,284 rows), labor stats, and state incomes into Cassandra/MongoDB, powering Kibana dashboards for county/state affordability analysis.
Women in STEM Evidence
Surfaced state-level disparities (e.g., women’s STEM share ≈29%) and pay gaps with choropleths and ranked gaps, giving policy stakeholders targeted levers for equity programs.
Better Minority-Class Sentiment
A hybrid TF-IDF + MPQA lexicon model improved precision/F1 on rare sentiment classes versus TF-IDF alone—useful for customer feedback signals where extremes are scarce.
Failure Risk Detection
On AI4I, class-balanced Random Forest achieved strong recall on failures, supporting proactive maintenance scheduling to reduce unplanned downtime.
Academic Credentials & Executive Certifications in Data Strategy and Governance
Get in Touch
Open to collaborations, advisory work, and leadership roles in data & AI.