Multimodal AI Systems Engineer

SpaceXSI; Palo Alto (CA); In loco; USD 160.000 - 250.000; Tempo pieno

Vantaggi offerti da questo lavoro

Equity

Comprehensive medical, vision, and牙?

401(k) retirement plan

Disability insurance (short & long‑‑)

Life insurance

Employee discounts

Descrizione del lavoro

SpaceXSI in Palo Alto is seeking a Model Engineer for Multimodal Understanding to push toward superhuman capabilities across image, video, audio, and text, spanning data curation, training, and deployment.

You will collaborate with pre-training, post-training, reasoning, data, and product teams to build scalable, end-to-end systems and interactive demos that operate in real time across modalities.

Competenze

• Production-quality code experience in Python, PyTorch, JAX, Rust, or C++.
• Expert-level Python with experience in PyTorch / JAX / XLA.
• Experience building or optimizing large-scale distributed ML systems.
• Strong data pipeline design for multimodal datasets.

Mansioni

• Design and optimize large-scale distributed systems for multimodal pre-training and inference.
• Develop high-throughput pipelines for data acquisition, preprocessing, filtering, and loading.
• Advance multimodal capabilities including spatial-temporal processing and cross-modal reasoning.
• Drive data quality studies and scalable pipelines for trillion-parameter models.
• Create evaluation frameworks and metrics for real-world usage and interactive dynamics.
• Experiment with algorithms, scaling, and hardware/software co-design for state-of-the-art performance.
• Build research tooling, prototypes, and end-to-end demos for rapid iteration.
• Collaborate across pre-training, post-training, reasoning, data, applied, and product teams.

Conoscenze

Python

PyTorch

JAX

Rust

C++

Strumenti

Spark

Ray

Kubernetes

Descrizione del lavoro

SpaceXSI in Palo Alto is seeking a Model Engineer for Multimodal Understanding to push toward superhuman capabilities across image, video, audio, and text, spanning data curation, training, and deployment.

You will collaborate with pre-training, post-training, reasoning, data, and product teams to build scalable, end-to-end systems and interactive demos that operate in real time across modalities.