Staff Machine Learning Engineer (Models)

Who are we?

Aarki is an AI-driven company specializing in mobile advertising solutions designed to fuel revenue growth. We leverage AI to discover audiences in a privacy-first environment through trillions of contextual bidding signals and proprietary behavioral models. Our audience engagement platform includes creative strategy and execution. We handle 5 million mobile ad requests per second from over 10 billion devices, driving performance for both publishers and brands. We are headquartered in San Francisco, CA, with a global presence across the United States, EMEA, and APAC.

Role Overview

We are seeking a Staff Machine Learning Engineer to build, deploy, and test machine learning models that power real-time advertising use cases including bidding, ranking, pacing, and fraud. This role focuses on delivering low-latency, high-throughput models into production with a focus on accuracy, stability, and explainability.

You will work at the intersection of machine learning and real-time bidding, partnering with our engineering team to serve millions of predictions per second in latency-sensitive environments. The ideal candidate brings strong systems programming experience and a production-first mindset.

What will you do?

• Own the design, development, and operation of end-to-end machine learning models, including model training, evaluation, and deployment.
• Partner with engineering to productize models and ensure reliable deployments.
• Optimize inference performance across the stack, including latency and calibration issues.
• Design, execute, and analyze A/B tests and clearly communicate results.
• Perform root-cause analyses on model behavior, identify causes of prediction changes, and implement preventative measures.
• Design and implement new ML features across our model training and inference pipelines.
• Optimize our model architectures for a combination of inference speed and accuracy.

What are we looking for?

• 8+ years of hands-on experience building and operating production machine learning systems
• Strong Python and Spark for large-scale processing (on-prem/YARN environments preferred)
• Proficiency in at least one systems language (e.g., C++, Java, Rust, Go) and strong experience with Python for ML development.
• Experience with data preparation and model training on large-scale data sets.
• Professional experience with deep learning frameworks.
• Familiarity with model serialization for serving, e.g., ONNX.
• Experience working with on-prem deployments of open source tools including Spark, ClickHouse, and Redash is a plus.

Nice-to-Have

• Exposure to online inference systems, gRPC/REST model endpoints, or streaming features (Kafka/Flink).
• Ad-tech familiarity: auction dynamics, pacing, fraud signals, creative personalization.