Oded — Enable Medicine (petabyte-scale spatial biology platform)
Yue — EM / Pathos / Tempus AI (foundation models, 1000 H200s)

Backed by Y Combinator

Our Vision

We build the missing data layer for the next generation of medicine — transforming sparse experimental data into complete predictions through computational biology.

Strand AI · Confidential

What we can do for Modularity Bio

In Silico Screening

  • Receptor-ligand prediction — screen 300+ receptors computationally before mouse studies
  • Signaling pathway modeling — predict JAK2 outputs from receptor configurations
  • Expression optimization — model protein ratios for multi-gene constructs

Infrastructure & Optimization

  • Inference optimization — we took CZI's VariantFormer from ~$93 to $1.50/sample
    63x faster · zero accuracy loss
  • GPU infrastructure — deep expertise running large-scale molecular ML workloads
    H200s · B200s

*We can work within your VPC or run on Strand AI compute and deliver the results.

Strand AI · Confidential

The Problem We Solve

Without Strand

  • Test 300 receptors → 300 mouse experiments
  • Months of iteration
  • High cost per candidate

With Strand

  • Train on your existing paired data
  • Screen computationally, validate top hits in vivo
  • 10-100x fewer animal experiments
Strand AI · Confidential

How It Works

  1. You share paired data — receptor sequences + measured signaling outputs
  2. We build a predictive model — trained on your proprietary data
  3. Screen in silico — rank your 300 receptors by predicted efficacy
  4. Validate top candidates — only run mouse studies on the best hits

Your data stays yours. Models can run in your infrastructure.

Strand AI · Confidential