I'm passionate about applying Deep Learning & Artificial Intelligence to scientific domains, with a focus on Cheminformatics and Bioinformatics. Now I am a CTO, Co-Founder @ Deep MedChem and exploring innovative AI applications in drug discovery and medical chemistry.
- Chemical Embeddings Search Engine – Custom implemented vector database with 10B+ isometric embeddings of molecules trained using novel AI approach. The UI is written in React/Typescript, Vite, Express and Tailwind CSS. Read paper preprint
- DOI Reference Extractor @ references.mireklzicar.com – Enter a DOI to extract and download all references in your preferred citation format
- Ketcher Docker @ ketcher.mireklzicar.com – A production-ready Vite-based application that embeds the Ketcher chemical structure editor with full iframe communication support.
- Issue Duration Labeller – Github Marketplace – Github Action for automatic labelling of issues based on time taken
- LineIndex – LineIndex provides lightning-fast random access to lines in large text files through efficient indexing. It's a pip package designed to handle very large files where you need to frequently access specific lines without reading the entire file.
- ESP and RESP partial charges – Streamlit app of computation of near DFT-quality ESP and RESP partial charges
- Modeller – Streamlit app allowing users to train custom AI models for ADMET property prediction and run inference at scale unmatched by competitors
- Spanning Trees Search – Python project leveraging matrix methods in graph theory to efficiently enumerate spanning trees.
- Particle Swarm Optimization – C# project for visualisation of particle swarm optimization.
- Time Resolved X-ray Diffraction Crystallography – Work for ELI Beamlines under research group TREX at Department of Structural Dynamics