Senior Data Scientist

Nathan Aguirre

Senior Data Scientist for personalization, recommenders, and customer behavior ML.

I work with product, engineering, and analytics teams to turn messy customer and product questions into models, pipelines, experiments, and dashboards that hold up after launch.

Open to Senior DS and Staff-scope roles - Select consulting

Discuss a full-time role Consulting inquiry LinkedIn

Experience

8+ years in applied data science

Systems

Production ML, pipelines, schemas, dashboards, monitoring

Problems

Personalization, churn, LTV, experimentation, risk

Ownership

Early-team tooling, standards, and senior IC execution

Selected work

Selected work across streaming, ecommerce, and insurance.

A few examples of applied data science work: what was needed, what I built, and how it was used.

Crunchyroll

Personalization and retention ML for global streaming

At Crunchyroll, I was the first data scientist on a growing data science function. My work covered recommendations, retention, customer value, and product measurement, usually starting from loosely defined product or marketing questions and turning them into pipelines, event definitions, dashboards, and maintained models.

Context: A global streaming product needed better discovery, retention, and customer-value signals while the data science function was still being built.
Built: Recommendation models, churn/LTV and user-behavior models, subtitle-informed LLM/RAG recommendation work, event schemas, PySpark/Python/SQL pipelines, KPI dashboards, and maintenance paths.
Measured: Campaign experiments, A/B testing workflows, model monitoring, KPI dashboards, and maintenance checks.
Result: 2x+ improvement in recommendation diversity/relevance, support for double-digit churn reduction, and hundreds of thousands of monthly cancellation events prevented.
Role: First data scientist on the team; helped establish tooling, event schemas, experimentation workflows, and recurring project patterns.

Nordstrom

Product discovery and customer activation models

At Nordstrom, my work sat close to the ecommerce experience: product discovery, catalog personalization, add-on recommendations, and customer activation. The common thread was using customer and product data to make shopping flows more relevant, not just building offline models.

Context: Ecommerce teams needed practical ways to use catalog and customer-behavior data in discovery, activation, and add-on recommendation workflows.
Built: Topic modeling, clustering, recommender models, gradient boosting, survival analysis, and exploratory NLP/computer-vision prototypes over customer and product data.
Measured: Customer-facing personalization and activation workflows tied to product and ecommerce outcomes.
Result: Contributed to catalog personalization, add-on recommendations, and activation workflows with measurable ecommerce impact.
Role: Worked close to product and analytics teams to connect models to shopping flows.

Allstate

Forecasting and road safety analytics

At Allstate, I worked on applied forecasting and road-risk analytics problems where noisy behavioral, sales, and geospatial data needed to become useful decision inputs.

Context: Extract usable forecasting and road-risk signals from noisy automotive sales and telematics data.
Built: ARIMA-style forecasting inputs, statistical risk methods, geospatial analysis, and telematics-derived road safety signals.
Measured: Validated road safety interventions and translated behavioral data into risk insights.
Result: Built forecasting inputs and road-risk methods used for stock forecasting and intervention validation.
Role: Turned forecasting and road-safety questions into usable analytical methods.

Foundations

NASA summers and chemical engineering training.

Before industry data science, I spent five summers at NASA: one at Kennedy Space Center on shuttle tile repair work, and four at Ames Research Center across clustering, visualization, and biochemical fuel cell projects. I later earned an MS in Chemical and Biological Engineering from Northwestern. That early work still shows up in how I approach data science: clarify the measurement problem, make assumptions visible, and build models that can be tested and maintained after launch.

Expertise

Problem areas I work in.

Grouped by the work people usually need done, not by every tool I have used.

Personalization and recommender systems

Collaborative filtering, factorization machines, ranking systems, product and content discovery, topic modeling, clustering, and language-model-assisted recommendation approaches.

Customer behavior modeling

Churn prediction, LTV forecasting, activation modeling, retention campaigns, survival analysis, segmentation, and campaign targeting.

Experimentation and product measurement

A/B testing pipelines, KPI dashboards, campaign evaluation, model impact measurement, and clear readouts for product and business partners.

Applied ML systems

PySpark, Python, SQL, Airflow, Databricks, MLflow, cloud deployment, monitoring, on-call support, and model maintenance.

Risk and anomaly modeling

Anomaly detection, behavioral risk signals, and operational model monitoring.

Forecasting, time-series, and spatial analysis

Sales forecasting, ARIMA-style forecasting, telematics analysis, spatial risk methods, and intervention validation.

Operating style

How I tend to work.

I am usually most useful where the model is only part of the job: defining the data, building the path to launch, measuring the result, and helping the team keep it running.

Early-team ownership

I have been the first data science hire on a team, setting up tooling, schemas, workflows, and reusable model infrastructure.

Build past the notebook

I am most useful when the work includes pipelines, dashboards, monitoring, and maintenance, not only offline model results.

Translate between teams

I work across product, marketing, engineering, and analytics so model outputs land in decisions people actually make.

Lead while staying hands-on

I have led small project teams while staying close to the modeling and delivery work.

Consulting

Focused ML and measurement support.

Consulting stays secondary to full-time roles, but I am open to scoped work where the problem fits my experience.

Available for focused ML and measurement projects.

I occasionally take on scoped projects where a team needs experienced help with applied ML, especially around personalization, recommender systems, churn and retention, experimentation, and customer behavior modeling.

Start a consulting conversation

Recommender system review
Churn, LTV, retention, and activation modeling
A/B testing and measurement design
Applied ML system review
Data science planning for early-stage teams
Short-term senior DS advising

Contact

Let's talk.

I am open to Senior Data Scientist and Staff-scope roles, especially around personalization, recommender systems, customer behavior, experimentation, and applied ML systems. I also take on select consulting projects when the scope is narrow and useful.

Discuss a full-time role Consulting inquiry LinkedIn