Earth Observation AI Research Scientist (Remote)

hum.ai; United States; Remote; USD 150,000 - 210,000; Full time

Job summary

hum.ai in San Francisco or remote is seeking an AI Research Scientist to own the model at the center of our system. You will guide the data ingestion, experiment with models, and collaborate with the Chief Scientist and engineers to deploy breakthroughs for customers and the world.

You'll lead architecture decisions, explore data semantics, define success metrics, and publish research on Earth observation foundation models, representing our work at conferences and in the community.

Qualifications

• PhD or equivalent experience in AI research with strong publication record.
• 2+ years of industry experience shipping AI systems.
• Experience designing models for unconventional data modalities beyond text.

Responsibilities

• Own the architecture. Evaluate our current approach, understand the reasoning, and adjust where the evidence supports.
• Figure out what the data knows by applying state-of-the-art techniques to uncover semantics in large datasets.
• Define success by helping decide the metrics we measure rather than chasing benchmarks.
• Go from small to large: prove ideas with fast experiments and scale winners to petabyte-scale training.
• Shape the integration: collaborate with the agent team on model invocation, outputs, and product usefulness.
• Plan the path: set milestones, roadmaps, and resourcing with leadership, spotting risks early.
• Share what you learn: publish research on earth observation foundation models and present at conferences.

Skills

AI systems development

AI research understanding

Education

PhD in computer science, engineering or related field

Equivalent experience

Tools

Python

PyTorch

Job description

hum.ai in San Francisco or remote is seeking an AI Research Scientist to own the model at the center of our system. You will guide the data ingestion, experiment with models, and collaborate with the Chief Scientist and engineers to deploy breakthroughs for customers and the world.

You'll lead architecture decisions, explore data semantics, define success metrics, and publish research on Earth observation foundation models, representing our work at conferences and in the community.