Teaching

Teaching

Courses, workshops, written material and the people I've supervised. Most of it is public, because a course nobody can look at is just a line on a CV.

I teach in both directions: computation to biologists, and biology to computer scientists. It is the same lesson from either side — the hard part is almost never the idea, it's that nobody ever sat down and showed anyone how to run the thing. Standing in the middle of that gap for long enough turns out to be a skill in itself.

In practice that has meant a three-day deep learning course for people who had never written a training loop, a crash course in molecular biology for people who had never seen a ribosome, university practicals, conference tutorials, hackathons, and a general-audience talk about protein structure delivered in a tent.

Courses and workshops

  1. 2024

    Deep Learning for Genomics

    Three-day course, University of Malta

    Convolutional and recurrent models for biological sequence data, taught from first principles to a room of biologists who had not written a training loop before, and who had one working by the end of it. Seven notebooks, from k-mers and a classical classifier through to transfer learning and interpreting a trained network — all public, all still runnable. It finished with a hackathon on real miRNA binding data, scored on a held-out set nobody had seen.

  2. 2023

    Explainability for Deep Neural Networks

    MALTAomics Summer School, University of Malta

    A hands-on session on getting a model to account for itself — saliency, attribution, and the gap between an importance score and an explanation. Notebooks are public and still run.

  3. 2021

    Inside the black box: explainable deep learning for biological data

    Tutorial, ECCB 2021

    A conference tutorial on interpreting sequence models, co-taught with the rest of the group. The materials have outlived the conference.

  4. 2020 – 2021

    Introduction to Bioinformatics

    Teaching assistant, Faculty of Medicine, Masaryk University

    Practicals for a medical-faculty bioinformatics course: the command line, data preprocessing, and the first programming most of the students had ever done. I wrote and published the practical materials.

Written for anyone

  1. ongoing

    The Missing Skills

    Project-based tutorials on the computational skills a science degree leaves out — version control, the shell, and making your work runnable by someone who isn't you. Written because the bottleneck in computational biology is rarely the idea; more often nobody ever showed anyone how to run the thing.

  2. 2021 – 2022

    Biology Crash Course

    Molecular biology from the ground up, written for computer scientists who need enough of it to be dangerous. The mirror image of The Missing Skills, and between them the two directions I spend most of my teaching in.

Supervision

  1. 2025 – present

    Master's thesis supervision

    Max Delbrück Center, Berlin

    Supervising a master's student on predicting translation efficiency from an mRNA's own sequence — working up from handcrafted features to pretrained foundation models, and asking at each step whether the extra machinery actually earns the complexity.

  2. 2023 – 2025

    Master's thesis co-supervision

    University of Malta

    Co-supervised a master's thesis on identifying miRNA targets with deep learning, from problem statement through to a defended thesis.

  3. 2023 – 2024

    Bachelor internships

    University of Malta

    Mentored three bachelor interns working on bias in genomic datasets — which is to say, on the reason the miRBench project existed at all.

Events organised

  1. 2022 – 2023

    Hackathons on miRNA binding and Genomic Benchmarks

    CEITEC and University of Malta

    Co-organised hackathons where the point was that everybody left with something running. Good practice for the thing I care about most: getting a method out of one person's notebook.

  2. 2023

    International workshop on similarity search

    Masaryk University

    Co-organised a workshop on machine-learning-based similarity search, bringing the structure-search and the sequence-model people into one room.

Talks and outreach

  1. 2026

    GenBenchQC: automated quality control of genomic sequence benchmarks reveals widespread biases in deep learning datasets

    EMBO | EMBL Symposium: AI and biology, EMBL Heidelberg

    A selected talk on auditing the datasets the field trains on, in a session with Oded Regev and Peter Koo. The short version: a benchmark nobody has checked is a benchmark a model can pass for the wrong reason.

  2. 2026

    GenBenchQC: automated quality control of genomic sequence benchmarks reveals widespread biases in deep learning datasets

    Short talk, RECOMB-Seq 2026, Thessaloniki

    The same work at the RECOMB satellite on biological sequence analysis, selected from the submitted abstracts for a slot in the main programme.

  3. 2024

    Digital revolution of proteins

    Researchers' Night, Brno

    Protein structure prediction, explained to a general audience including a large number of children, in a tent.

  4. 2024

    miRBench: benchmark datasets for microRNA binding site prediction

    Poster, ECCB 2024, Turku

  5. 2024

    miRBench: benchmark datasets for microRNA binding site prediction

    Poster, Microsymposium on RNA Biology, Vienna

  6. 2023

    Understanding miRNA binding site prediction through deep learning models

    Data Science Platform Seminars, University of Malta

  7. 2023

    Attribution-based interpretation of deep learning models

    Molecular Medicine Seminars, CEITEC, Masaryk University

  8. 2022

    Computational modelling of miRNA targets

    XXI. Meeting of Biochemists and Molecular Biologists, Brno

  9. 2022

    Genomic Benchmarks: a collection of datasets for genomic sequence classification

    Poster, ENBIK 2022 — selected in the top five