Why Prometheus?

The tools of scientific discovery have never been more accessible. Protein language models can predict molecular properties in seconds. Machine learning can sift through telescope data faster than any human. Large language models can help write, analyze, and iterate on research.

But most of these tools sit idle — used for chatbots and image generation instead of pushing the boundaries of what we know.

Prometheus changes that.

We point AI at real scientific problems: searching for new antimicrobial peptides in extreme environments, hunting for exoplanets in telescope data, and exploring the boundaries of what’s possible with citizen science.

What We’re Building

🧬 AMPHunter

Antimicrobial resistance kills over 1 million people per year. We use Meta’s ESM-2 protein language model to scan metagenomic sequences from extreme environments — hot springs, glaciers, deep-sea vents — for novel antimicrobial peptide candidates.

🪐 ExoHunter

Building machine learning pipelines to detect exoplanet transits in TESS satellite data. Combining feature engineering, gradient boosting, and neural networks.

🌌 ExoExplorer

An interactive dashboard for exploring 5,700+ confirmed exoplanets, complete with habitable zone visualization and AI-generated alien ecosystems.

Open Science

Everything is open source. Every sequence, every model, every line of code. If we find something interesting, the world gets to see it — not just a paper behind a paywall.

⚠️ Important caveat: All our results are computational predictions. Nothing has been experimentally validated. We’re transparent about our methods and their limitations.

What’s Next

We’re currently running our AMP pipeline across multiple extreme biomes and refining our exoplanet detection models. Follow along on our blog or check our GitHub for the latest code.

---

Prometheus — Bringing fire to science.