Senior Machine Learning Engineer, Moloco NEXT

Moloco•Menlo Park, CA

8d•Onsite

About The Position

Moloco is a company that builds advanced AI advertising solutions. Our mission is to democratize access to sophisticated AI technology. We are seeking a Senior Machine Learning Engineer to join our Moloco NEXT team, which focuses on performance advertising. In this role, you will be responsible for the CTR/CVR prediction models within a real-time bidding system, making critical decisions in under 100ms. This is an opportunity to own a significant part of our ML stack, identify new data signals, run rapid experimentation cycles, and contribute to the technical direction of the platform. We value engineers who can handle ambiguity, embrace the less glamorous aspects of ML engineering like data quality and consistency, and collaborate effectively with other teams.

Requirements

5+ years of machine learning experience.
Proven track record of shipping production-grade models in business-critical environments.
Experience with data analysis.
Experience working on large-scale prediction or decisioning systems (e.g., CTR/CVR, ranking, recommendation, personalization).
Proficiency in writing code in Python.

Nice To Haves

Experience using LLMs as agents or feature extractors, particularly in evaluation, experiment debugging, or signal-discovery contexts.

Responsibilities

Own a production CTR/CVR prediction model end-to-end, including modeling, evaluation, feature pipelines, online experimentation, and post-launch operations.
Identify and integrate new data sources and derive contextual features to improve model performance.
Design offline evaluation strategies, conduct online A/B testing, and analyze results rapidly.
Diagnose and resolve offline-online divergences in model performance.
Develop agentic tooling to automate experiment debugging and signal discovery workflows.
Set technical direction for the ML stack, deciding on future priorities and bets.
Collaborate with data and pipeline teams to ensure effective signal integration.
Focus on data quality instrumentation, train/serve consistency in feature pipelines, and detailed evaluation slicing to identify failure modes.