An ML Engineer with a strong background in developing and applying machine learning models to complex biological data. Experienced in the end-to-end development of computational tools for 3D bio-imaging and multi-modal data analysis. A proficient Python developer focused on leveraging advanced AI techniques, including Transformers and CNNs, to solve challenges in oncology and neuroscience.
Career Journey & Academic Background
Developed AI models for oncology tissue profiling by integrating multi-modal spatialomics data. Utilized Transformers, ResNet architectures, and generative models (VAEs) for tissue morphology analysis. Analyzed single-nucleotide variants (SNVs) and copy number alterations in prostate cancer using data mining methods.
Led the development of scalable computational workflows to analyze large-scale, multimodal neuroscience imaging data. Designed and deployed user-facing Python-based GUIs, creating human-in-the-loop pipelines to accelerate scientific discovery. Led a technical team of 4 scientists and mentored 2 interns, driving project execution and fostering innovation.
Research in microfluidics, neuronal cell culture, and computational pipeline development.
1st Class Honors; IMechE Best Project Award.
Computational Biology Research & Development
An AI-based image registration tool for aligning brain slices in the olfactory system to 3D volumes.
Python • Deep Learning • Image Processing • 3D RegistrationBuilt a computational framework for in-vivo to ex-vivo registration of individual neurons.
Python • Computer Vision • Neural Registration • MicroscopyBuilt models for automated tissue profiling in oncology using multi-modal data.
Transformers • ResNet • VAEs • Spatial AnalysisA pipeline for processing and analysing in-situ sequencing images.
Python • Image Processing • Bioinformatics • SequencingCreated scalable pipelines for sequencing and imaging data using tools like Snakemake.
Snakemake • Python • HPC • Cloud ComputingInteractive visualization tools for biological data exploration including Embedly for high-dimensional embeddings and Spatial-Embedly for dual-window spatial analysis.
Python • Dash • Plotly • UMAP • Interactive VisualizationExpertise in AI/ML, Bioinformatics & Software Engineering
Explore biological data through interactive demos
Explore latent embeddings of prostate tissue patches from the PESO dataset. Each point represents different tissue types. Hover over points to see the original histology images.
Explore the relationship between spatial coordinates and embedding space with synchronized highlighting across both views.
Classic MNIST digit visualization demonstrating UMAP dimensionality reduction techniques.
Registration of ex-vivo to in-vivo of single neurons in the brain. Built using Napari, a human-in-the-loop GUI with a point cloud registration algorithm in the background.
Recognition for Research Excellence & Innovation
2016
2013
University of Aberdeen
Education & Mentorship in Data Science & Bioinformatics
University of Oxford
2024
University of Oxford
2025
University of Oxford
2018-2020
University of Oxford
2018
US Patent (2024)
Patent: 11,931,735
For a complete list of publications, visit my Google Scholar profile
Email: soitu.cristi@gmail.com
Location: Oxford, United Kingdom
Available for opportunities in computational biology and AI/ML roles