We are seeking a Machine Learning Engineer to join our Campaign Delivery team. This role is central to designing and deploying predictive models and recommendation systems that power how campaigns are matched to the right users at the right time. Youll partner closely with Data Engineers, Product Managers, and Campaign Operations to build ML solutions that directly improve delivery performance, advertiser outcomes, and platform efficiency.

A key focus will be developing and iterating on models that optimize campaign targeting, pacing, and conversion prediction across delivery infrastructure. Youll work with large-scale behavioral and event data to train, evaluate, and deploy models that operate in real-time production environments.

A major initiative will be automating the conversion review process. Today, conversion approvals involve manual effort to distinguish legitimate installs and post-install events from fraudulent or low-quality traffic. You will design and deploy ML models that score conversions in real time, flag suspicious patterns, and progressively automate approval decisions—reducing manual review volume while protecting advertiser spend and publisher trust. This work sits at the intersection of anomaly detection, classification, and production ML, and will directly shape how quality assurance scales across the delivery pipeline.

This is a hands-on modeling role. You will build, train, evaluate, and deploy machine learning models end-to-end—not integrate third-party AI APIs or wrap LLM services.

RESPONSIBILITIES

• Design and deploy ML models to automate the conversion review and approval process—scoring conversions for legitimacy and flagging fraudulent or low-quality traffic patterns.

• Build and maintain fraud detection and traffic quality models that identify install fraud, click injection, device farms, SDK spoofing, and other invalid traffic signals across the delivery pipeline.

• Design, build, and deploy machine learning models for campaign targeting, bid optimization, conversion prediction, and user-value scoring.

• Develop and maintain recommendation systems that match campaigns to high-value user segments based on behavioral signals, contextual data, and historical performance—ranging from gradient-boosted approaches on structured features to embedding-based and deep learning methods as complexity warrants.

• Build predictive models for campaign pacing, budget allocation, and performance forecasting to maximize ROI (CPI, CPA, ROAS, LTV).

• Collaborate with Data Engineering to design and maintain feature pipelines that feed real-time and batch ML models, including feature store development.

• Partner with Product and Engineering to integrate ML capabilities into production systems with a focus on reliability, latency, and scalability.

• Design and execute rigorous A/B tests and offline experiments to validate model performance and quantify business impact, including power analysis, confidence intervals, and guardrail metrics.

• Monitor model performance in production, implement drift detection, and establish retraining cadences to maintain accuracy over time.

• Contribute to the development of MLOps infrastructure, including model versioning, deployment pipelines, experiment tracking, and model registries using tools such as MLflow or similar platforms.

• Stay current with advancements in applied ML, recommendation systems, fraud detection, and adtech-specific modeling techniques.

REQUIRED EXPERIENCE & QUALIFICATIONS

• 5+ years of experience in machine learning engineering or applied data science, with a strong track record of shipping models to production environments.

• Strong proficiency in Python and hands-on experience with tabular ML frameworks such as scikit-learn, XGBoost, and/or LightGBM.

• Demonstrated experience building recommendation systems, ranking models, click-through rate (CTR) prediction, conversion rate models, or similar predictive systems at scale.

• Experience building classification or anomaly detection models—ideally in fraud detection, traffic quality, conversion validation, or similar trust-and-safety domains.

• Experience with feature engineering, feature stores, and data pipelines using tools like Spark, Airflow, or dbt; familiarity with experiment tracking and model lifecycle management tools such as MLflow.

• Solid understanding of model evaluation methodology, experimentation design, and A/B testing with statistical rigor.

• Experience deploying and serving models in production via REST APIs, containerized services, or serverless architectures (AWS SageMaker, Lambda, ECS, or similar).

• Familiarity with cloud infrastructure (AWS preferred) and data warehouses (Redshift, Snowflake, or similar).

• Strong communication skills with the ability to translate complex technical concepts into business narratives, in spoken and written English.

PREFERRED EXPERIENCE

• Experience in adtech, performance marketing, or mobile user acquisition—familiarity with KPIs such as CPI, CPA, ROAS, eCPM, LTV, and install-to-event conversion rates.

• Experience with mobile install fraud detection techniques: click injection, device farms, SDK spoofing, click flooding, or attribution manipulation.

• Experience with deep learning frameworks (PyTorch, TensorFlow) for recommendation systems, embedding-based models, or representation learning on user-behavioral data.

• Experience with online learning approaches, multi-armed bandits, or contextual bandits for real-time decisioning.

• Familiarity with bid optimization, real-time bidding (RTB), or programmatic advertising systems.

• Experience with large-scale user behavioral data and event-stream processing.

• Exposure to causal inference or uplift modeling techniques for campaign optimization.

• Comfort leveraging AI-assisted development tools to accelerate engineering workflows.

PROFILE

• Curious mindset with a drive to ask questions and uncover opportunities.

• Ownership mentality and bias toward shipping and iterating.

• Comfortable collaborating across product, engineering, and data teams.