Primum non nocere, first do no harm, is medicine's oldest commitment. It is also, literally, what our name means. Everything else about how we work follows from this.
Ainnocence is a next-generation biotech company with an AI platform designed to minimize drug discovery risk and cost. Our system runs lightning-fast virtual screening and multi-objective pharmacological optimization at scale that was infeasible a decade ago.
With the computational capacity to screen up to 10 billion protein sequences or chemical compounds within hours, Ainnocence reduces discovery time and cost while achieving exceptionally high wet-lab hit rates. The platform makes it tractable to tackle targets that were previously undruggable or structurally intractable.
But capacity is table stakes. The reason the company is named Ainnocence is that we design molecules that go into people. That obligation shapes how we work, and it is what distinguishes us from a general-purpose ML platform rented out to pharma.
"We envision AI as a catalyst for innovation, driving a better world through transformative advancements in healthcare. Our mission is to make precision medicine accessible, not aspirational."
· Ainnocence vision statement
Four commitments that distinguish us from a general-purpose model dropped into a biotech. Each is a design decision we can be audited on, not a slogan.
Downstream patient impact is a design constraint, not an afterthought. Every target that enters the platform is checked against known liabilities, off-target risks, and dual-use concerns before our model generates a single sequence.
Partners get the real performance of the platform, hits, misses, and the campaigns that didn't converge. Our internal track record is an honest statistic, not a filtered highlight reel.
Credibility depends on tests that resist easy wins. AINN-P1's leave-program-out AUC (0.81 vs 0.68 for finetuned ESM2) was reported on a benchmark we constructed ourselves to be difficult, including for us.
The scientists, clinicians, and partners who move a molecule forward hold the responsibility for the decision. Our platform is a decision support system, not an autonomous one, and we think that's a feature, not a limitation.
“We report what we found, not what we hoped. We publish failures alongside successes, disclose uncertainty honestly, and correct the record when we are wrong. Metrics are earned, not marketed.”
Meet the scientists, engineers, and chemists behind this work on the team page.
Our data room is open under CDA. Pick a target, we'll share wet-lab data and freedom-to-operate analysis within one week.
Request the data room →