Zenbase AI
S24Continuous prompt optimization for LLMs from DSPy core contributors
100x faster scientific evidence synthesis
Zenbase helps developers focus on programming by automating prompt engineering and model selection. We’re core contributors of Stanford NLP’s DSPy, the #1 LLM optimization framework used by Meta, Microsoft, Google, among many others. But it's hard to use outside of academia. We're building the production version.
We enable research that was previously impossible—systematic reviews of 20,000+ papers in weeks with 98% accuracy. We collaborate with researchers at Stanford, Harvard, and Johns Hopkins to ask the big questions they've been forced to scale down. Pharma companies get complete evidence awareness in weeks — not years. AI labs get the same reviews as RL environments with expert-validated reasoning traces—scientific synthesis as the next frontier after math and coding benchmarks. We're former DSPy core contributors who built the production infrastructure to make total evidence awareness actually work at scale.
Zenbase shifted from automating LLM prompt optimization for developers (essentially a developer tools/NLP infrastructure focus) to building a scientific evidence synthesis platform for researchers and pharma—entirely different product and market.