Peloton Interactive, Inc. seeks Senior Machine Learning Engineer in New York, NY:

Job Duties: Build and improve machine learning pipelines that power Peloton’s content recommendations. Research and apply best-in-class machine learning techniques for recommender systems. Evaluate, implement, and improve machine learning models. Run A/B tests and experiments and analyze the results in collaboration with the product analysts. Productionize, deploy, and monitor machine learning models and services. Collaborate and work closely with the platform teams to leverage their tools and infrastructure to rapidly iterate ideas that drive delightful, personalized experiences for millions of users. Mentor and guide junior engineers on technical best practices, design patterns, and project execution, fostering professional growth within the team. Serve as a subject matter expert (SME) for Infrastructure as Code (IaC), driving the adoption of state-of-the-art tooling and contributing to codebase optimizations. Conduct thorough and timely code reviews to ensure adherence to high engineering standards, security protocols, performance requirements, and team best practices. Take primary on-call responsibility for a critical service or functional area, demonstrating the ability to independently identify, scope, and implement service enhancements. Exhibit strong product sense by collaborating cross-functionally with Product Managers and User Research teams to generate hypotheses, inform user studies, and design features that directly align with business outcomes and core member needs. Architect and implement complex, end-to-end Machine Learning system architectures that are SOTA, scalable, reliable, and maintainable, often involving multiple services and platforms. Identify areas and implement solutions for reduction in operational expenditure (OpEx) for ML services by optimizing inference latency, implementing efficient resource provisioning strategies, and continually refining infrastructure scaling policies to maximize cost-efficiency. Taking a leading role in the recruitment process, including structuring interviews, evaluating candidates, and making hiring. Create and maintain comprehensive documentation for ML systems, best practices, and internal libraries to ensure knowledge transfer across the organization. Telecommuting is an option. Some travel to Peloton offices may be required. Salary: $200,740-$271,000.

Minimum Requirements: Master’s degree (or its foreign degree equivalent) in Computer Science, Engineering (any field), or a related quantitative discipline, and four (4) years of experience in the job offered or in any occupation in related field.

Special Skill Requirements: Requires at least three (3) years of experience with the following: (1) Architecting and implementing large-scale distributed data processing pipelines using Apache Spark on Amazon EMR to perform ETL operations, feature engineering, and data transformations on multi-terabyte datasets derived from user workout telemetry and content metadata; (2) Developing real-time event-driven architectures using Apache Kafka for high-velocity data ingestion and leveraging Amazon Kinesis Data Firehose to reliably stream transformed telemetry into data lakes or warehouse destinations; (3) Conducting deep-dive diagnostics on Transformer, DLRM, and XGBoost architectures to identify systemic failures like training-serving skew and gradient vanishing across multi-terabyte datasets. Architecting sophisticated feature engineering strategies and implementing high-throughput pipelines for sub-millisecond online inference using Feast to ensure model reliability; (4) Implementing vector-based similarity searches and complex aggregations, enabling high-performance candidate retrieval for personalized content discovery using Amazon Opensearch or Neo4j graphs while implementing a low-latency and highly scalable system; (5) Deep understanding of Transformer-based recommendation architectures using self-attention, Two-Tower models, DLRM, Xgboost and Encoder Decoder architectures. Ability to select ideal algorithms, address cold start and exploration/exploitation trade-offs, and build scalable training pipelines using Airflow to tune models for deeper user engagement; (6) Optimizing machine learning models for production using the ONNX format and also experience in deploying them via NVIDIA Triton Inference Server to maximize hardware utilization, minimize inference latency, ability to discover and fix bottlenecks and support multi-framework model serving; (7) Architecting and managing scalable ML microservices using Kubernetes, MLFlow, Redis, DynamoDB, GitOps (Argo CD), and full-stack observability (Datadog) to ensure high-availability. This includes conducting comprehensive cost audits and implementing optimization strategies to significantly reduce OpEx while maintaining high-performance; (8) Designing and executing online A/B tests and offline experiments (evaluating precision/recall, MAP@K, and AUC) in collaboration with product analysts; performing in-depth statistical analysis on interaction datasets to validate model performance and ensure features directly align with business outcomes and member needs.
Requires at least one (1) year of experience with the following: (1) Leveraging Amazon Bedrock to architect and fine-tune LLMs like GPT OSS and GPT-4o mini for member data processing; designing and deploying high-performance LLM workflows for batch and real-time inference to deliver scalable, low-latency generative AI features that deepen Peloton member engagement; and (2) Overseeing the development of features with high ambiguity and providing technical leadership/mentorship to junior engineers, including conducting thorough code reviews, defining engineering best practices, and architecting end-to-end ML system designs that ensure security, performance, and maintainability. Telecommuting is an option. Some travel to Peloton offices may be required.

Submit a resume with references to: Req. # L22-129859 via the Peloton Careers webpage: https://www.onepeloton.com/careers or by email at: onlinejobpostings@onepeloton.com.