Senior Deep Learning Engineer, Cosmo 3D Spatial

NVIDIA
Santa Clara
Workplace: OnsiteFull timeUSD 224,000 - 356,500 annuallyFunction: Data Science & Machine LearningExperience: 12+ yearsSkills: ["Communication","Partnering","Evaluation methodology","Rigor","Collaboration"]

Build and own the data engine behind Cosmos 3D spatial reasoning, from sourcing and curating 2D/3D vision-language training corpora to designing 3D-grounded annotation and auto-labeling pipelines. Drive end-to-end data quality and training verification across large multi-node GPU clusters, and maintain a 3D/spatial evaluation suite with benchmark traceability. Partner with Cosmos Lab researchers to translate hypotheses into scalable ablations and shipped model improvements.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
NVIDIA
NVIDIA
2 hours ago

Senior Deep Learning Engineer, Cosmo 3D Spatial

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 2 hours agoStatus: Live

Job Summary

Build and own the data engine behind Cosmos 3D spatial reasoning, from sourcing and curating 2D/3D vision-language training corpora to designing 3D-grounded annotation and auto-labeling pipelines. Drive end-to-end data quality and training verification across large multi-node GPU clusters, and maintain a 3D/spatial evaluation suite with benchmark traceability. Partner with Cosmos Lab researchers to translate hypotheses into scalable ablations and shipped model improvements.
Location: Santa Clara
Workplace: Onsite
Employment Type: Full time
Job Function: Data Science & Machine Learning
Seniority: Mid level

Key Responsibilities

  • •Own the 3D data engine: source, curate, filter, and balance large-scale image/video corpora into vision-language training data for spatial understanding.
  • •Build 3D-grounded annotation and auto-labeling pipelines and supervision for camera-relative 3D boxes, reference frames, correspondence, and question answering.
  • •Manage end-to-end data quality and coverage with deduplication, automated quality scoring, coverage analysis, and balanced sampling strategies.
  • •Verify end-to-end model training by running and validating pre-training and supervised fine-tuning pipelines and diagnosing data/checkpoint regressions.
  • •Build and operate the 3D/spatial evaluation suite and benchmarks, and partner with researchers to translate hypotheses into dataset and ablation experiments.

Pay and Benefits

Salary: USD 224,000 - 356,500 annually
Equity and Bonus:Equity

Key Requirements

  • •MS or PhD in Computer Science, Electrical/Computer Engineering, Robotics, or related field, or equivalent experience.
  • •12+ years building deep learning systems in Python with PyTorch or JAX on Linux.
  • •Deep expertise in 3D computer vision and practical experience producing and validating 3D ground truth at scale.
  • •Hands-on experience building vision-language training data and evaluations that improve visual grounding and reasoning.
  • •Experience running and validating large model training on multi-GPU, multi-node clusters with distributed/sharding strategies (e.g., data, tensor, pipeline parallelism or FSDP).
Experience:12+ yearsDeep learningComputer vision3D visionMultimodal learning
Education:
Skills:CommunicationPartneringEvaluation methodologyRigorCollaboration
Tech Stack:PythonPyTorchJAXLinuxRaySparkSlurmFSDPVLMEvalKit

Company Brief

NVIDIA
Designs and manufactures GPUs, AI accelerators, and system-on-chip products for gaming, data centers, professional visualization, and automotive markets, enabling advanced graphics, AI, and high-performance computing solutions worldwide.
Industry: Electronics Manufacturing
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Public Company
Valuation: Public Company (Market Cap in USD)
Funding: IPO / Publicly Listed
Headquarters: Santa Clara, United States
Founded: 1993
Glassdoor
Glassdoor: 4.3
WebsiteLinkedInGlassdoor