Osseus GTM brain

Osseus develops biomedical datasets, evaluations, reinforcement learning environments and post-trained models for AI labs and life-sciences applications.

category

Biomedical AI training and evaluation

market profile

  • text: Medical and biological AI developers need domain-specific data and evaluation as they improve model capability and safety.
  • text: The lab positions its work across training data and post-training for health AI and pharmaceutical applications.

personas

  • title: Frontier model research lead
    pains:
    • General-purpose benchmarks may miss medical or biological failure modes.
    • Needs realistic expert-informed evaluations and training environments for specialized capabilities.
  • title: Pharma AI research director
    pains:
    • General models may lack sufficient performance on specialized biological tasks.
    • Acquiring and curating domain data and measuring progress require scarce scientific expertise.
  • title: Health AI product leader
    pains:
    • Needs to demonstrate model reliability in clinical contexts before deployment.
    • Must coordinate model development with researchers and clinicians.

icp

  • text: Frontier AI labs building or evaluating models for medicine and biology, including organizations with internal post-training teams.
  • text: Pharma, biotech and healthcare AI teams pursuing domain-specific model applications; account size and geography are not stated.

message pillars

  • text: Treat biomedical data as a research challenge.
  • text: Build training and measurement infrastructure for future models.
  • text: Adapt models to health and pharmaceutical tasks.

objections

  • objection: How can we trust the scientific quality of your data and evaluations?
    response: Request a scoped example, expert review process and held-out evaluation design before committing.
  • objection: Can you support our specific biomedical task?
    response: Discuss the task, data rights and measurable success criteria directly with the research team.

competitors

  • name: Centaur AI
    website: https://centaur.ai/
    why: Provides expert biomedical labeling and model evaluation.
  • name: Grayde
    website: https://grayde.ai/
    why: Offers clinician and scientist data, evaluations and training for health and life-science AI.
  • name: CellType
    website: https://www.celltype.com/tasks
    why: Offers reinforcement learning environments for drug discovery.

influencers

  • name: Ayush Patel
    url: https://www.linkedin.com/in/ayush-patel4
    why: Co-founder posting about Osseus and its research direction.

keywords

  • medical AI training datasets
  • biomedical model evaluation
  • healthcare AI reinforcement learning environments
  • clinical AI post training
  • biological AI benchmarks
  • medical model safety evaluation

voice

  • Research-oriented
  • Ambitious
  • Collaborative

signals

  • text: A model lab is expanding into medical or biological reasoning and needs domain evaluations.
  • text: A pharma or health AI team is planning domain-specific model post-training.

pricing

model: unknown

motion

sales_led

channels

  • research publications
  • LinkedIn
  • X
  • YC launch

confidence

medium

Founders

  • name: Anish Mariathasan
    title: Founder/CEO
    role: CEO
    bio: Co-Founder at Osseus (YC S26). Prev. co-founded a compliance AI startup. ML research at a Harvard-HKU lab on computer vision for cancer research. MEng in Biomedical Engineering from Imperial College London, dropped Oxford PhD in AI Safety (for Health). Awarded academic scholarships for both.
  • name: Ayush Patel
    title: Founder/CTO
    role: CTO
    bio: Co-Founder at Osseus (YC S26). 2x intern at Apple, where he built an agentic QA system. MEng in Computer Science from Imperial College London.

Back to YC companies