Senior Data Scientist, AdTech & Media Buy Optimization

Launch Potato; Valparaíso; Hybrid; CLP 137,525,000 - 196,464,000; Full time

Job summary

Launch Potato is seeking a senior data scientist to own end-to-end ML models that optimize insurance marketplace decisions and media buying. You will identify data, build production-grade models, validate performance, and monitor drift in a fast-paced, remote-first environment across multiple countries.

You will work with an ML engineer on deployment and scaling, handling complex feature sets, late labels, and evolving objectives while ensuring measurable impact and ROAS improvement.

Qualifications

• 5+ years hands-on data science with measurable business impact.
• Models used to drive paid media decisions: spend, bidding or ad serving; attribution and reporting not sufficient.
• Built and explained model architectures, loss functions and training data considerations.
• Handled late labels, sparse features or drift in dynamic environments.
• End-to-end ownership from data discovery to deployment and monitoring.

Responsibilities

• Frame business problems and identify data to address them.
• Build and validate production ML models for insurance marketplace decisions.
• Deploy and monitor models; collaborate with ML engineers on the production side.
• Own data pipelines and validation to support scalable bidding and targeting.
• Iterate models as market conditions evolve and ensure measurable impact.

Skills

Data science

Machine learning

Modeling

Python

SQL

Education

Bachelor's or higher in a quantitative field

Tools

AWS SageMaker

AWS ECS

S3

Athena

Python

Job description

Launch Potato is seeking a senior data scientist to own end-to-end ML models that optimize insurance marketplace decisions and media buying. You will identify data, build production-grade models, validate performance, and monitor drift in a fast-paced, remote-first environment across multiple countries.

You will work with an ML engineer on deployment and scaling, handling complex feature sets, late labels, and evolving objectives while ensuring measurable impact and ROAS improvement.