Data Engineer

INVID•San Juan, PR

7h•Hybrid

About The Position

What sets INVID apart is our collaborative and flexible work environment. We encourage our team to raise the bar in everything they do while maintaining a healthy work-life balance. With our hybrid work model, team members thrive both in the office and remotely. We foster a culture of mutual respect, autonomy, and accountability, where your voice matters and your growth is supported. From structured career paths and paid professional development to access to industry events, we’re committed to your success. Join us at INVID, where innovation meets support, and together we deliver excellence. Job Description We are hiring a Data Engineer to build the data infrastructure powering our predictive analytics initiative. You will create the pipelines that turn raw vessel tracking data into training datasets for ML models. The core challenge: we have rich behavioral data (vessel positions, AIS gaps, ship-to-ship transfers, spoofing events) but limited labeled outcomes (confirmed violations, detentions, seizures). You will build pipelines that create usable training data through proxy labels, data joins, and outcome correlation.

Requirements

4+ years data engineering experience
Strong SQL skills, including complex joins across large datasets
Experience with Spark, Airflow, or equivalent distributed processing frameworks
Python for data processing and pipeline orchestration
AWS experience
Understanding of ML training data requirements

Nice To Haves

Experience with geospatial data (PostGIS, H3, spatial joins)
Maritime, defense, or intelligence domain experience
Experience with data labeling infrastructure or weak supervision
Familiarity with real-time streaming data systems

Responsibilities

Build labeling pipelines that join behavioral events to outcome data (sanctions designations, flag changes, detentions)
Implement proxy labeling strategies that create training signal from observable outcomes
Build weak supervision infrastructure to combine multiple noisy labeling rules
Create and maintain ML training datasets at scale
Build data validation and quality monitoring systems
Implement versioning for reproducible model training
Integrate LRIT position data for prediction validation
Build pipelines that compare predicted locations against actual LRIT reports
Create feedback loops that improve model accuracy over time
Scale data infrastructure as models and data sources grow