Software Engineering PMTS - ML Engineer

About the Team:
Digital Enterprise Technology (DET) at Salesforce builds the internal engines that power our business. The Machine Learning & Innovation (MLI) team is the applied science and infrastructure strike-force responsible for "Foundry," our unified ML/AI Platform. We operate as "Customer Zero," building the high-velocity, real-time streaming feature platforms and recommendation models that power internal products like the Events App and enterprise-wide personalization.
Role Overview:
As a Senior Staff Data Engineer specializing in real-time telemetry, you will build the high-velocity streaming pipelines that feed our online feature platforms. You will tackle the hardest problems in stream processing—managing late-arriving data, maintaining stateful aggregations, and integrating complex third-party datasets—to ensure our machine learning models react to user behavior instantly.
What You'll Do:
Design and implement robust upstream event telemetry ingestion pipelines, ensuring high-fidelity data capture from source systems.
Build and maintain complex stream processing applications (e.g., Flink, Spark Streaming) to compute real-time feature aggregations.
Integrate diverse third-party datasets (e.g., demographic and firmographic data) into unified streaming feature definitions.
Directly resolve critical engineering bottlenecks related to event ingestion into the Zipline feature platform.
Define and deploy live feature contracts and monitoring systems to guarantee the freshness and correctness of online features.
Collaborate with ML engineers to ensure streaming telemetry schemas perfectly align with the requirements of real-time inference models.
What You Bring:
15+ years of experience in data engineering, with a strong specialization in high-throughput stream processing and real-time analytics.
Deep expertise in streaming frameworks (Apache Flink, Apache Spark Streaming) and event brokers (Apache Kafka, AWS Kinesis).
Strong proficiency in Python, Java, or Scala.
Proven experience working with real-time feature stores (Zipline, Feast) and managing the complexities of stateful stream computation.
Familiarity with AWS infrastructure and streaming data architectures.
A strong focus on system resilience, low-latency processing, and exact-once/at-least-once delivery guarantees.

Employment type
Full time