About the role
TL;DR
Lead engineering initiatives to scale ML innovation at Upstart.
- •As a Principal Machine Learning Engineer at Upstart, you will lead engineering initiatives that turn high-impact modeling needs into scalable, reusable infrastructure.
- •You will work at the intersection of applied ML and platform engineering, collaborating closely with Research Scientists, Data Scientists, and ML Platform Engineers.
- •Key Responsibilities Scale ML innovation by building tools, infrastructure, and workflows that dramatically improve the speed and reliability of model development.
- •Work backward from modeling needs to design systems that directly unlock gains in accuracy, efficiency, and scientific productivity.
- •Explore new algorithms and methodologies for our machine learning models and develop tooling to support them.
- •Improve the entire ML lifecycle—from data readiness and feature development through training, evaluation, serving, and monitoring.
- •Automate and standardize operational workflows, enabling scientists to focus on high-leverage modeling and analysis rather than manual pipelines.
- •Requirements Strong grasp of ML fundamentals and statistics.
- •Deep knowledge of the entire modeling lifecycle
- •from data preparation to training and deployment to production.
- •Experience with building a unified embeddings platform for training, serving, and managing representations at scale.
- •Experience with streamlining feature engineering pipelines to reduce manual steps and deliver new signals quickly.
- •Experience with developing automated continuous-learning systems that handle data refresh, retraining, evaluation, and drift monitoring with minimal manual effort.
Required skills
TensorFlowPyTorchscikit-learnPandasNumPySparkAirflowdbtPostgreSQLMySQLMongoDBRedisAWSGoogle CloudAzure
Nice-to-have skills
KerasTableauPower BILookerSnowflakeDatabricksFivetranDagsterPrefectApache Kafka
Domain expertise
fintech
Tech stack
MLflowKubeflowVertex AIONNXJAXTensorFlowPyTorchscikit-learnKerasComputer VisionNLPLLMsPandasNumPySparkAirflowdbtTableauPower BILookerSnowflakeDatabricksFivetranDagsterPrefectApache KafkaApache FlinkPostgreSQLMySQLMongoDB