Building a new age of digital biology.

AI / ML / data engineer with roots in bioinformatics. Belgrade.

I build domain-specific AI/ML for the life sciences — narrowly scoped models and agents that take the complexity of living systems seriously. Constant learner, challenge enthusiast.

Co-Founder & Systems Architect at Applied Scientific Intelligence (ASI) — a bootstrapped lab ending hidden cardiac toxicity. ASI pairs prediction models with reasoning agents to build a Virtual Heart that catches toxicity before it kills a drug program.

1. Alexandria — SOTA scientific-literature agent, ranked #1 on major literature benchmarks (LitQA3, FigQA2 & TableQA2), built with NVIDIA.
2. CardioSafe — multi-channel cardiac safety model (hERG, Naᵥ1.5, Caᵥ1.2, IKs); AUC 0.919, beating published baselines.
3. RefusalBench — open benchmark auditing how often frontier models refuse legitimate biological research.

Try it out