Postdoctoral Researcher — Edge Vision & Foundation AI

NASA Jet Propulsion Laboratory (JPL); United States; On-site; USD 75,000 - 92,000; Full time

Benefits offered by this job

Health benefits

Retirement plans

Paid time off

Flexible schedule

Job summary

NASA Jet Propulsion Laboratory in Pasadena, California invites applications for a Postdoctoral scholar to advance edge Vision-Language Models and Foundation models. The successful candidate will deploy state-of-the-art AI on embedded devices, fine-tune models (LoRA/SFT), and develop combined vision heads for segmentation and open vocabulary detection.

Publications and conference presentations are expected. The appointee will collaborate with a JPL advisor, with potential for multi-year renewal,

Qualifications

• Ph.D. in Computer Science, AI, CV, Robotics or closely related field.
• Research in Vision-Language Models, Foundation models, or embedded AI.
• Experience deploying AI on embedded platforms (NVIDIA NX/AGX).
• Hands-on with LoRA and SFT fine-tuning and multi-head vision models (segmentation, YOLO).
• Experience with evaluation/benchmark pipelines including LLM-driven evaluation.
• History of peer-reviewed publications.

Responsibilities

• Conduct research in edge Vision Language Models (VLMs) and Foundation models.
• Deploy foundation models to embedded platforms (e.g., Nvidia NX, AGX).
• Fine-tune foundation models using LoRA and SFT.
• Develop and integrate multiple vision heads for segmentation, object recognition, open vocabulary detection.
• Design, implement, and run evaluation and benchmark pipelines, including LLM-driven methods.
• Publish results in peer-reviewed journals and present at conferences.

Skills

Vision-Language Models

Foundation models

Embedded AI

LoRA fine-tuning

SFT fine-tuning

YOLO

Evaluation pipelines

Education

PhD in CS/AI/CV/Robotics

Tools

NVIDIA NX

NVIDIA AGX

Job description

NASA Jet Propulsion Laboratory in Pasadena, California invites applications for a Postdoctoral scholar to advance edge Vision-Language Models and Foundation models. The successful candidate will deploy state-of-the-art AI on embedded devices, fine-tune models (LoRA/SFT), and develop combined vision heads for segmentation and open vocabulary detection.

Publications and conference presentations are expected. The appointee will collaborate with a JPL advisor, with potential for multi-year renewal,