About Alkira Bio
Built by chemists. Designed around the real shape of early discovery work.
We are a small team of medicinal chemists and ML engineers in Melbourne. We set out to remove the ranking bottleneck from early-stage hit-to-lead programs.
Why we started
The problem is not synthesis. It is deciding what to synthesise first.
Medicinal chemistry programs have always produced more candidate compounds than any team can test. The choice of which 20 from a 300-compound library to prioritize for synthesis drives the entire program timeline.
That decision was made by hand, with spreadsheets and competing expert intuitions, for most of the history of the field. We built Alkira Bio to make it faster, more reproducible, and more transparent, without taking the chemist out of the loop.
Our scoring engine trains on real binding, toxicity, and synthesizability outcomes from active programs. Not historical public data alone. The result is a model that reflects the constraints of real discovery work, not idealized academic benchmarks.
Alkira is a computational research tool for early-stage discovery prioritization. It does not constitute a clinical or regulatory decision system. Scores inform synthesis queues, not patient treatment decisions.
Team
The people behind the platform
Medicinal chemist by training. Spent years building hit-to-lead series at a Melbourne drug discovery company before co-founding Alkira Bio in 2023. The scoring methodology reflects what bench teams actually need from a ranking tool, not what looks good on an academic benchmark.
ML engineer focused on molecular property prediction. Developed graph neural network architectures for ADMET prediction in a computational chemistry research context before co-founding Alkira Bio. Leads model architecture and infrastructure.
PhD medicinal chemist. Practiced synthetic chemistry at a small-molecule biotech before moving into computational approaches. Joined Alkira Bio in 2024 to lead scoring model validation and chemist-facing product design.
Backing
Building with patient capital
We took angel backing in 2025 to accelerate model training on real program data. Our investors have operational backgrounds in biotech and drug discovery. They are not in a rush to sell a product before the science is right.
We are not chasing a growth curve at the expense of accuracy. The scoring engine ships when it meets our internal precision thresholds, not when a funding timeline demands it.
Try the scoring engine on your own compound series.
14-day free trial. No credit card. Cancel any time.