I work with small biotechs to extend the computational capabilities behind their science, rather than outsourcing the thinking. Quick answers when that's all you need; durable capability your team can own when that's what moves the needle.
From rational protein engineering to AI-driven de-novo design — and the cloud infrastructure to run it at scale. I take projects end-to-end or plug into your team where you need depth.
Rational engineering of mutants, fusions and fragments — solubility, stability, glycosylation, IP position. Increasingly built around AI methods: AlphaFold, RFdiffusion, ProteinMPNN, ESM-based scoring.
Reverse vaccinology pipelines, antigen and epitope selection, HLA and strain coverage, population centroid design. Tissue-specific bicodon optimization and UTR selection for proprietary mRNA platforms.
Bespoke model development where off-the-shelf tools fall short — from fine-tuning protein language models to building scoring functions, prediction pipelines and agentic workflows around your data.
Heterogeneous biomedical data — PubMed, MeSH, drug and target databases, patent full-text — woven into queryable graphs and pipelines. Optional setup of reproducible AWS environments for compute-heavy work.
Translating bioinformatic results into biomedical decisions. IP background and freedom-to-operate searches, custom literature reviews, drug repositioning via protein/drug/disease networks.
A specific question, a few days, a concrete deliverable — a low-stakes way to see how we work together before any larger commitment. Even a short engagement is done transparently and documented: nothing is a black box. You can take the process in-house if and when it makes sense for you, or just take the result; both are fine.
Outsource everything and you own the output and nothing else — and you start from zero on the next program. Extend your own capability and every project after the first is faster, the expertise stays in your team, and you're not dependent on anyone to run your own science. Investors notice the difference too: owned capability reads as a defensible platform, not a one-off result.
Twenty-five years in pre-clinical biotech across infectious disease, oncology, mRNA, antigen design and drug repositioning.
AlphaFold, RFdiffusion, ProteinMPNN, ESM, and bespoke models — paired with classical structural and statistical methods where they outperform.
You work with me, not with an account manager. NDAs welcome, IP stays with you, no offshore subcontracting.
If a project is better done in-house, with a different vendor, or with a wet-lab partner first, I'll say so.
Molecular geneticist (PhD, University of Vienna). 25+ years across startups and established biotech — protein, peptide and RNA engineering, computational vaccinology, and AI-driven de-novo protein design. Former Associate Director at BioNTech R&D Austria (2023–2025).
When I started, “bioinformatics” meant BLAST searches, hand-curated motifs and a lot of waiting on jobs. Today, I can sketch a binder geometry in the morning, generate sequences with diffusion models by noon, and have folded structures by afternoon. The fundamentals — biology, careful evaluation, knowing what a result actually means — matter more than ever.
Sodatana is how I bring that combination — deep domain knowledge plus current AI craft — to projects that need it. If that sounds like yours, let's talk.