ML Foundation Engineer
Remote / Hybrid
About the Role
We’re looking for an ML Foundation Engineer to join Falcon’s Data Team and take ownership of the end-to-end machine learning foundation behind Falcon’s core ad-server product.
This role is responsible for building and operating production-grade ML systems for ad ranking and decisioning, from data validation and modeling through real-time deployment and continuous learning.
Hands-on experience building ML systems in an ad-tech environment is a big advantage for this role.
Your Role on the Data Team
As part of Falcon’s Data Team, you will:
Own the ML foundation powering Falcon’s ad ranking and optimization
Design and deploy ML systems operating under high scale environment
Work closely with the Performance & Data Projects Lead to connect experimentation and ML decisioning
What You’ll Do
ML System Ownership
Translate business objectives into ML optimization goals
Design and build ML ad ranking based on predictive models
Own the full ML pipeline
Build and maintain a research and evaluation framework for model iteration and improvement
Enable offline testing and comparison of modeling approaches before production deployment
What We’re Looking For
Required Experience & Skills
Proven experience as an ML Engineer building end-to-end ML pipelines, preferably in an ad-tech environment
Strong understanding of ad-tech optimization dynamics (e.g. attribution, delayed feedback, exploration vs. exploitation, pacing)
Strong foundations in machine learning, applied statistics, and mathematical reasoning
Strong data engineering skills in Python
Experience working in Snowflake environment is a plus
Strong ownership mindset, ability to work independently, and collaborate closely within a data team
What Success Looks Like
ML-based ranking outperforms existing baselines in production
Models are stable, monitored, and improve continuously
Falcon’s ML foundation scales with network growth
A robust offline research and evaluation framework exists to iterate on models and test improvements before production
A clear production evaluation and rollout framework is in place to safely test and compare new model versions in live traffic
A scalable signal research pipeline enables rapid exploration and validation of new features and data sources