Corporate Workshops

Build a small tool for a recurring task at work with AI, and learn to test whether it works correctly — with step-by-step guidance.

• Start with a tool that summarizes meeting notes, using a workspace prepared for you.
• Describe a task to the AI, try the tool it creates and check its results with examples.
• Show what your tool does, explain where it fails and identify when to ask an engineer for help.

Who it is for: Operations, administrative and domain specialists.

Schedule: Day 1: 5 h 45 min; day 2: 6 h, plus breaks and lunch

Preparation and scope: We prepare the workspace and exercises. Use example data unless another data policy is agreed. You will demonstrate a small tool in the workshop; a software professional should review it before it is used in day-to-day work.

Duration: 2 days

Format: Step-by-step practical workshop

Prerequisites: No programming experience required. Bring a recurring task to discuss; guided project options are available.

Turn scattered AI experiments into a phased plan with clear leadership commitments and measures of progress.

• Write a personal leadership commitment with a date and a measure of progress.
• Compare training and adoption approaches using organizational cases, then choose an approach for your context.
• Draft a phased plan, decide when to continue or stop and how to use time freed by AI.

Who it is for: Engineering and business leaders responsible for AI adoption.

Schedule: 6 hours, plus breaks and lunch

Preparation and scope: Bring one adoption challenge and a description of what your organization has already tried. Participants write throughout the session and leave with a draft plan to refine and implement in their organization.

Duration: 1 day

Format: Working session

Prerequisites: A leadership role and an organizational adoption challenge to work on.

Help a colleague reproduce a research result without relying on your machine or undocumented setup knowledge.

• Try to reproduce a result, identify missing setup instructions and record exact software and data versions.
• Add checks on the results and verify or isolate unchecked changes, including AI-generated changes.
• Have a colleague reproduce the result and document remaining work with owners and next steps.

Who it is for: Researchers whose code, models or analyses must be reproduced by others.

Schedule: Day 1: 5 h 45 min; day 2: 6 h, plus breaks and lunch

Preparation and scope: Agree on one feasible result and repository before the session. If your code is confidential or too large for the workshop, use sanitized data or the provided practice repository. The reproduction test covers the selected result and documented conditions.

Duration: 2 days

Format: Hands-on workshop

Prerequisites: Ability to run and make basic edits to research scripts. Use a repository scoped before the workshop or a provided practice repository.

Practice numerical computing with attention to reproducibility, numerical stability and accuracy.

• Floating-point arithmetic & numerical errors
• Algorithm stability & conditioning
• Linear algebra best practices
• Reproducibility frameworks

Duration: 2 days

Format: Hands-on workshop

Prerequisites: Basic programming knowledge

Measure where Julia code spends time and compare ways to improve it using optimization and parallel computing.

• Performance profiling & benchmarking
• Multi-threading & parallelism
• GPU computing with CUDA.jl
• Type stability & compiler optimization

Duration: 3 days

Format: Intensive hands-on lab

Prerequisites: Julia basics, linear algebra

Practice using AI assistants to write and revise scientific software, then check the generated code with tests.

• Prompt engineering for code generation & debugging
• AI-assisted refactoring & documentation
• Testing AI-generated scientific code
• Working with language-model assistants and checking their output
• Limitations & validation strategies

Duration: 2 days

Format: Interactive workshop with hands-on exercises

Prerequisites: Basic programming knowledge, familiarity with Python or Julia

Combine machine learning with physical models and differential equations, then evaluate predictions and uncertainty.

• Physics-informed neural networks (PINNs)
• Neural ODEs & differential equation solvers
• Uncertainty quantification
• Model validation & benchmarking

Duration: 4 days

Format: Intensive hands-on training

Prerequisites: Basic machine learning, calculus and linear algebra

Practice sharing and reviewing research code with Git, including how to combine changes and resolve conflicts.

• Advanced Git workflows (branching, merging, rebasing)
• Code review best practices & pull request etiquette
• Collaborative development standards
• Automated testing and release checks for scientific software (CI/CD)

Duration: 1 day

Format: Workshop with practical exercises

Prerequisites: Basic familiarity with version control

Practice reorganizing research prototypes so a team can run, test and maintain them. RSE means research software engineering.

• Modular architecture design
• Application programming interfaces (APIs) and deployment
• Data pipeline engineering
• Monitoring software behavior and diagnosing failures

Duration: 3 days

Format: Workshop with live refactoring

Prerequisites: Research code to reorganize for use and maintenance by a team

Compare ways to organize software components, define interfaces and manage dependencies as a project grows.

• Domain-specific language (DSL) design
• Monorepo architecture & tooling
• Dependency inversion & decoupling
• System integration patterns

Duration: 4 days

Format: Architecture workshop + case studies

Prerequisites: Software engineering experience

Write tests for scientific calculations and practice test-driven development: write a failing test, implement the behavior and improve the code.

• Test-driven development (TDD) methodology
• Testing strategies (unit, integration, end-to-end)
• Test coverage metrics and quality standards
• Testing scientific computations and numerical accuracy
• Mocking, fixtures, and test data management
• Continuous testing in CI/CD pipelines

Duration: 2 days

Format: Workshop with test-driven development exercises

Prerequisites: Working knowledge of at least one programming language

Represent images as numerical arrays and write scripts to repeat an analysis across many images. Check the results against examples.

• NumPy array operations & image representation
• Scikit-image for filtering and segmentation
• Batch processing strategies & file I/O
• Pandas integration for quantitative results
• Quality control & validation metrics

Duration: 3 days

Format: Hands-on workshop

Prerequisites: Basic Python knowledge

Build analysis tools for Napari, a multidimensional image viewer for Python, and add graphical controls that colleagues can use.

• Napari architecture & plugin ecosystem
• MagicGUI for rapid interface development
• Widget design patterns & user workflows
• Testing & documentation standards
• Publishing Python packages on PyPI and managing versions

Duration: 2 days

Format: Hands-on workshop

Prerequisites: Python proficiency, basic image analysis experience

Practice training and evaluating microscopy models. Examine annotation quality, prediction errors and the limits of applying a model to new images.

• Ground truth annotation strategies & quality control
• Model architecture selection (StarDist, Cellpose, U-Net)
• Cloud-based training pipelines (GPU optimization)
• Validation metrics & error analysis
• Running trained models and monitoring their predictions

Duration: 4 days

Format: Hands-on workshop

Prerequisites: Python, NumPy, basic machine learning concepts

Compare and implement advanced deep-learning approaches for scientific images, including learning from unlabeled data and combining different data types.

• Custom CNN and transformer architectures
• Self-supervised and unsupervised learning for imaging data
• Multimodal data fusion (e.g., imaging + sequencing)
• Advanced image segmentation techniques (e.g., panoptic segmentation)
• Generative models for image restoration and augmentation

Duration: 5 days

Format: Hands-on workshop

Prerequisites: Python, PyTorch, experience with deep learning models

Translate biological questions into mathematical models and explore their behavior through simulation and analysis.

• Ordinary differential equation (ODE) based modeling
• Parameter estimation and sensitivity analysis
• Bifurcation analysis and multistability
• Stochastic modeling of biological processes
• Network analysis of biological pathways

Duration: 4 days

Format: Hands-on workshop

Prerequisites: Python, calculus, linear algebra, basic biology concepts