Right, not right-sounding
A confident guess is not good enough for science. Every answer we serve traces back to a real source, with the provenance and citations to prove it, so researchers can check our work the way they check anyone's.
Building the foundation for people-centered, AI-accelerated research
Valency was founded on a simple conviction: the new era of AI-accelerated science is still, and will be, fundamentally a human-centered endeavor. We are a Berkeley-based team building the connective tissue between the world's research and the tools that reason over it.
What we believe
A few convictions shape everything we build. They are why Valency looks the way it does.
Science moves in hours. So do we.
A paper posted this morning should be findable this afternoon. We index new work as it appears, because a frontier you reach a quarter late is not the frontier.
The scientist stays in the loop
AI accelerates researchers. It does not replace them. Every product decision starts from that order, and we design for the person doing the science, not around them.
Leadership
We're a small team of scientists and builders who have spent their careers close to research, and to the infrastructure it runs on. Valency is led by enterprise founders with multiple exits.
Josh Bloom
CEO and Co-founder. Josh is an academic and serial entrepreneur. As an astronomy professor at UC Berkeley he has worked in the interface between time-domain and compact-object astrophysics. He received the Pierce Prize from the American Astronomical Society and is a former Hertz Foundation Fellow, Sloan Foundation Fellow, and a Junior Fellow at the Harvard Society. He holds a PhD from Caltech, MPhil from Cambridge, and AB from Harvard. Before founding Valency, his previous AI-for-customer-support company, wise.io, was acquired by GE.
Ryan Anderson
COO and Co-founder. Ryan leads operations and enterprise strategy for Valency. He was an Architect in Residence at IBM, IBM CTO for the Palo Alto Networks ecosystem, and senior director of R&D and Data Science at Connect Solutions. His IBM edge computing team was a founding member of the Linux Foundation's LF Edge, and he served on the LF Edge board. He holds an MBA from Cambridge University.
Karthik Ram
Head of Science. Karthik sets the direction for what the company builds across product, engineering, and marketing. His focus is on making AI actually useful for research: tools that meet scientists where they already work. He has founded and led national initiatives for open science and research software sustainability, and holds a PhD in Ecology and Evolution from UC Davis. He is a senior fellow at the Berkeley Institute for Data Science.
Matt Work
Chief Revenue Officer. Matt is a technology executive and entrepreneur who has built and scaled numerous software, data, and AI businesses. Before joining Valency, he co-founded Mechanical Orchard leading operations, strategy, and financing behind its mission to modernize legacy mainframe systems. Prior to that he was CEO of Discovery Mining, an e-discovery SaaS company acquired by Interwoven; VP of Strategy at Cognizant; and head of research at Alexa Internet.
Advisors
Our advisors have built and run the tools, archives, and platforms that research depends on, and study how science itself moves forward.
Fernando Pérez
Co-founder, Project Jupyter. Fernando is Professor of Statistics at UC Berkeley, Faculty Director of the Berkeley Institute for Data Science, co-faculty-director of the Schmidt Center for Data Science & Environment, and a faculty scientist at Lawrence Berkeley National Laboratory. He also created IPython, co-founded 2i2c, and was a founding NumFOCUS board member. His awards include the FSF Award for the Advancement of Free Software (2012), the ACM Software System Award, and the NASA Exceptional Public Service Medal.
Photo: Kayla Sim / UC Berkeley College of Computing, Data Science, and Society
Steinn Sigurðsson
Scientific Director, arXiv. Steinn is also Professor of Astronomy and Astrophysics at Penn State. His theoretical work spans compact-object binaries and black holes, dense stellar-system dynamics, gravitational-wave sources, and exoplanets. Elected a Fellow of the American Physical Society in 2021, he is an Honorary Member of the Aspen Center for Physics (trustee 2015–2021) and a scientific editor at The Astrophysical Journal.
James Evans
Professor of Sociology & Data Science, University of Chicago. James founded and directs Knowledge Lab and co-directs the Chicago Center for Computational Social Science. He also holds an appointment at the Santa Fe Institute. His work uses large-scale text, bibliometric, and now generative-AI methods to study how collectives think and how science and technology actually innovate. His honors include the ASA Star-Nelkin Award, an IBM Data Fellowship, and a Javits Fellowship, and in 2024 he received an NSF award to build LLM “virtual laboratory” models that predict scientific discovery.
Darren Haas
Founding member of Siri. Darren is a world-renowned engineer and operator. He built the original prototype backend for Siri, then led cloud engineering for Siri and Speech at Apple through 2016. He also co-founded Change.org. He went on to GE Digital as SVP of Platform Cloud Engineering (where he was the technical buyer behind GE’s acquisition of wise.io, Josh’s previous company), then to enterprise technology at AWS, and then co-founded Voltron Data.
How we work
Building infrastructure for science means holding ourselves to the standards of the people who use it.
Build with researchers, not just for them
We design alongside working scientists at universities and national labs, and ship against real workflows instead of imagined ones. Their feedback sets our roadmap.
Ship grounded
We would rather say "we don't know yet" than emit something plausible and wrong. Provenance and correctness come before polish, in the product and in how we talk about it.
Small team, wide ownership
We keep the team small and the ownership broad. Everyone talks to users, everyone touches the product, and good ideas win regardless of who has them.
Get in touch
Press and media: media@valency.io. Labs, institutions, and partnerships: labs@valency.io. Want to build this with us? See our open roles.