AI Researcher

hum.ai; San Francisco (CA); Hybrid; USD 120,000 - 160,000; Full time

Job summary

A cutting-edge AI startup is searching for an experienced AI Researcher eager to advance generative AI. This role requires a PhD and 5+ years of research experience, focusing on developing innovative models that harness earth observation. The ideal candidate will demonstrate proficiency in machine learning, have experience with large models, and excel in both research and publication. Competitive salary options and a potential for remote work for exceptional candidates are available.

Qualifications

• 5+ years of experience in machine learning research projects.
• Strong scientific understanding of generative AI.
• Experience in training large diffusion or transformer models.

Responsibilities

• Manage and execute a research agenda independently.
• Develop high-level proof-of-concept models.
• Assist in preparing and writing scientific papers.

Skills

Machine learning research

Deep learning

Generative AI

Scripting languages

Team collaboration

Education

PhD in computer science or related field

Tools

Python

PyTorch

Ray Trainer

Anyscale framework

Job description

AI Researcher

Location: San Francisco

About

Hum.ai is building planetary superintelligence. Backed by top funds, we’ve raised $10M+ and are now heads down building.

Join us at the cutting edge, where we’re scaling generative transformer diffusion models, designing next‑gen benchmarks, and engineering foundation models that go far beyond LLMs. You’ll be at the core of a moonshot journey to define what’s next in agentic AI and frontier model capabilities.

We are looking for an experienced AI Researcher who is eager to advance the frontier of AI, help us design, build, and scale end‑to‑end novel foundation models, and leverage their hands‑on experience implementing a wide range of pre‑training and post‑training models, including large foundation models (beyond just LLM fine‑tuning).

Who are we?

Hum is a seed‑funded startup on a mission to create positive impact through earth observation and AI. Founded at the University of Waterloo by a team of PhDs and engineers, we’re backed by some of the best AI and climate tech investors like HF0, Inovia Capital and Propeller Ventures, angels like James Tamplin (cofounder Firebase) and Sid Gorham (cofounder OpenTable, Granular), and partners like Amazon AWS and the United Nations.

What do we do?

We’re building multimodal foundation models for the natural world. We believe there’s more to the world than the internet + more to intelligence than memorizing the internet. Our models are trained on satellite remote sensing and real world ground truth data, and are used by our customers in nature conservation, carbon dioxide removal, and government to protect and positively impact our increasingly changing world. Our ultimate goal is to build AGI of the natural world.

The role will involve:

• Research & Design:

  • Deep understanding of current machine learning research.

  • Proven track record of generating new ideas or enhancing existing ones in machine learning, evidenced by first‑author publications or projects.

  • Contribute to research that uncovers the semantics of large datasets, with a focus on earth observation and remote sensing data.

  • Ability to independently manage and execute a research agenda, selecting impactful problems and conducting long‑term projects autonomously.

• Implementation:

  • Developing high‑level proof‑of‑concept (POC) models.

  • Experimenting with new state‑of‑the‑art techniques to surpass existing models.

  • Plan and execute innovative research and development to push the boundaries of current technology.

• Writing and Publication:

  • Assisting in the preparation and writing of technical and scientific papers for publication.

  • Demonstrated strong scientific communication/presentation skills.

Requirements

• PhD degree in computer science, engineering, a related field, or equivalent experience.

• Proven track record of successful machine learning research projects.

• 5+ years of experience.

• Strong scientific understanding of the field of generative AI.

  • Preference for experience training large diffusion or transformer models, especially on video or time series data.

• Preference for San Francisco or Waterloo, but we’re willing to consider remote for great candidates.

Nice to have

• Proficiency in scripting languages such as Python, Bash, or PowerShell.

• Demonstrated experience with deep learning and transformer models.

• Familiarity with designing, pre‑training, and fine‑tuning large models.

• Proficiency with Python, Ray Trainer, PyTorch, and Anyscale framework.

• Strong technical engineering skills.

• Previous experience in creating high‑performance implementations of deep learning algorithms.

• Past training of video or time‑series models.

• Team player, willing to undertake various tasks to support the team.

• Familiarity with cloud platforms such as AWS, GCP, or Azure.