ML Encoder Lead

About the role

Our client is seeking a senior hands-on ML Encoder Lead to design and train a shared customer representation model using large-scale longitudinal transaction, sales, and interaction data. This role will define the modeling objective, develop and evaluate encoder/embedding approaches, and determine whether the model provides meaningful downstream value for GenAI and analytics use cases.

Responsibilities

• Design and train encoder or embedding models using longitudinal customer/event data.
• Define pretraining objectives, modeling approach, evaluation design, and success criteria.
• Build inductive representations that can support customers/entities with limited history.
• Evaluate model performance using time-based splits, cold-start slices, leakage detection, drift, calibration, and transfer to downstream tasks.
• Build production-ready training pipelines, data contracts, model versioning, serving, monitoring, and reproducibility practices.
• Present findings, uncertainty, limitations, and go/no-go recommendations to senior stakeholders.

Requirements

• Hands-on experience personally training encoder or embedding models.
• Deep expertise in representation learning, self-supervised or contrastive learning, sequence/temporal modeling, transformers, GNNs, or recommender embeddings.
• Experience with large, sparse, longitudinal event data such as transactions, claims, clickstream, customer journeys, or engagement histories.
• Strong Python, PyTorch or JAX, SQL, distributed data processing, and cloud model training experience.
• Strong evaluation discipline and ability to identify whether embeddings create true incremental signal.
• Applicants must be authorized to work for ANY employer in the U.S. (no visa sponsorship available).

Preferred Skills

• Customer-360, behavioral embeddings, recommender systems, or foundation models over event data.
• Privacy, fairness, and re-identification risk experience.
• Publications, patents, or public applied work in representation learning.

Schedule

• 4-month contract position.
• Hybrid work arrangement: onsite at least 3 days per week in San Francisco, CA.
• West Coast-based candidates preferred; remote candidates must work Pacific Time hours.

Pay

$80.00–$100.00 / hour (USD).

Benefits

• Medical, Vision, and Dental Insurance Plans.
• 401k Retirement Fund.