Building the infrastructure for scarless regeneration.

AI/ML Architect and Founder specializing in multi-omic latent alignment, automated DBTL screening cycles, and Nextflow pipeline orchestration.

My Vision

Making Biology Engineerable

Drug discovery pipelines now rely on experimental work and manual target identification, success rate <10%.

My goal is to completely automate the biological Design-Build-Test-Learn cycle.

This requires a foundational data layer. My current work focuses on mapping multi-omic (transcriptomic and epigenomic) shifts into a unified computational latent space. By converting it to math coordinates, we can systematically model and execute virtual CRISPR perturbations.

This allows us to mathematically engineer cell states and accelerate the efficiency of therapeutic target identification.

News & Traction

Updates

  • Working on BioChemCPA — redesigning ChemCPA for embracing Biological Entities.[linkedin article][git repo]
  • GRN Consensus and reliable benchmarking.[linkedin article][git repo]
  • Updated and expanded StateZero platform — metabolomics, Reactome mapping, growth factors, and diffusion models for drug candidate selection.
  • Dissected SCENIC+ pipeline and extracted standalone GRN-inference core.[grn-core-source - git repo][article]
  • Initial concept for Cell Atlas — interactive cell-state transition mockup.[CellAtlas]
  • Started architecting StateZero Labs platform for developing regenerative therapies.[StateZero]
  • Built advanced segmentation CV pipeline for medical images — analysis of 20k+ murine ear images.
  • Deployed automated SAM-based workflow for medical image analysis for the lab, reducing processing time 10×.[sam - git repo]
  • Built cellular automata explaining origin of protein life — from entropy and initial gas state to template-directed self-assemblies. All personal questions on the origin of life answered; satisfied my curiosity.[article]
  • Started Undergraduate Research Fellow position at the Laboratory for Regenerative Medicine, Gdańsk Tech.

About

About Me

AI/ML Architect and Founder specializing in multi-omic latent alignment, automated DBTL screening cycles, and Nextflow pipeline orchestration.

I build Bio-MLOps infrastructure at StateZero Labs — systems that connect single-cell transcriptomics, virtual CRISPR screening, and chemical latent-space alignment into reproducible pipelines for target identification and cellular regeneration.

Core Infrastructure Stack

NextflowDockerPythonPyTorchScanpySCENICReactome

Current Deployment

Currently at the Max Delbrück Center (Bunina Lab), working on transcriptomics, GRNs and perturbation algorithms.

Experience

Roles

  • Visiting Computational Researcher

    Bunina Lab, Max Delbrück Center · Berlin, DE

    Jun 2026 — Oct 2026
  • Undergraduate Research Fellow

    Laboratory for Regenerative Medicine, Gdańsk University of Technology

    Oct 2025 — Present

Research

Research & Publications

  • BioChemCPA: Expanding ChemCPA Model to Biological Entities

    Abramau, P.

    Preprint · July 2026 · With apart research, CIC Secret Loyalties Hackathon

  • Linking CRISPR to Chemical Space — StateZero Labs Engine

    Abramau, P.

    2026 — currently in work

  • Investigating Mouse Ear Pinna as Regeneration Model

    Sachadyn P., Abramau P., et al.

    Preprint · 2026 — coming soon

Education

Academic Background

  • Internship — Computational Biology

    Max Delbrück Center for Molecular Medicine

    Jun — Oct 2026
  • B.Sc. Erasmus+ Biotechnology & AI

    University of Barcelona

    Feb — Jun 2026
  • B.Sc. Biotechnology

    Gdańsk University of Technology

    2024 — 2028

Projects / Pre-prints

Research Output

StateZero Platform

Virtual CRISPR engine for in-silico target prioritization — simulating perturbation outcomes across multi-omic latent spaces before wet-lab validation.

Automated Metabolic Parsing

LLM-driven pipeline integration for Reactome and GSEA — extracting structured pathway hypotheses from unstructured literature and experimental readouts.

Contact

Currently securing a remote Autumn infrastructure sprint.