Staff Research Engineer - Vision-Language Models for Autonomy

Rivian; Sacramento (CA); On-site; USD 265,000 - 331,000; Full time

Benefits offered by this job

401(k) Plan

Employee Stock Purchase Program

Job summary

Rivian is seeking a Staff Research Engineer to lead the VLM strategy within the Autonomy stack, focusing on training, shipping, and optimizing VLM models and extending multi-modal capabilities for new use cases.

You will own the end-to-end lifecycle of VLM model delivery, from data acquisition to deployment and feedback loops, collaborating with core Autonomy teams to translate requirements into robust ML entregables.

Qualifications

• 5+ years of professional experience scaling ML solutions.
• Hands-on VLM model training or fine-tuning using LoRA/QLoRA and RL alignment.
• Large-scale data mining techniques for long-tail distributions and rare cases.
• Zero-shot or few-shot classification with open vocabulary models.
• Strong Python and understanding of perception pipelines, benchmarking tools, and infrastructure.

Responsibilities

• Define, drive and execute the VLM model delivery roadmap including training, deployment and extension to multi-modalities.
• Design and deliver models for automated data mining, long-tail distributions, rare-edge-case detection and anomaly detection at scale.
• Establish rigorous evaluation/monitoring benchmarks and diagnose system-level anomalies.
• Collaborate with Autonomy teams to translate vehicle feature requirements into ML deliverables.
• Define system requirements and guide cross-functional alignment and trade-offs.

Skills

VLM training

Data mining

Zero-shot / few-shot

Python

System engineering

Performance optimization

Education

BS/MS/PhD in CS/Robotics/EE

Tools

TensorRT

Job description

Rivian is seeking a Staff Research Engineer to lead the VLM strategy within the Autonomy stack, focusing on training, shipping, and optimizing VLM models and extending multi-modal capabilities for new use cases.

You will own the end-to-end lifecycle of VLM model delivery, from data acquisition to deployment and feedback loops, collaborating with core Autonomy teams to translate requirements into robust ML entregables.