ADVANCED DATA SCHOOL: ML & AGENTIC CODING
A 5-day bootcamp focusing on advanced machine learning, agentic coding and rigorous evaluation.
TARGET AUDIENCE
Alumni of the introductory Data School (researchers, PhDs, doctors, analysts, government workers) seeking to master advanced concepts.
Developers and DBAs transitioning to data science who want hands-on practical application.
Managers and CEOs wanting hands-on understanding of agentic AI and modern data workflows, with some knowledge of Python or programming.
Prerequisites
Assumed Baseline Knowledge: basic knowledge of Python, Pandas, data manipulation, linear and logistic regression.
COURSE STRUCTURE
Day 1: Agentic Workflows and Data Processing
- Agentic Coding 101: Planning and context management
- Pandas data manipulation refresher
- SQL fundamentals: Joins and Window functions
- Planning and code verification in agentic workflows
Day 2: Feature Engineering and Trustworthy Baselines
- Feature engineering: missing values and encoding
- Training baseline models and proper data splitting
- Identifying and preventing data leakage
- Diagnostics: Bias-variance and cross-validation
- Model calibration and reliability
Day 3: Advanced Modelling and Explainability
- Algorithm comparison: LASSO vs. Gradient Boosting
- Hyperparameter tuning via agentic loops
- Explainable AI: SHAP values deep dive
- Interpretability vs. prediction power
- Evaluating structured data limitations
Day 4: Unstructured Data and App Production
- Extracting features from text via LLMs
- Signal vs. Noise Audit of LLM features
- Streamlit app prototyping
- Packaging models and SHAP results into applications
Day 5: Agent-Assisted Hackathon
We have prepared challenging data tasks where you will apply your newly acquired skills to build an end-to-end machine learning pilot. In practice, you will experience how agentic workflows can accelerate your coding, navigate their pitfalls, and uncover useful data insights in the process. You will conclude by presenting your findings, interactive app, and methodology to your colleagues.
AFTER THE BOOTCAMP, YOU WILL BE ABLE TO
- Use agentic coding tools (Antigravity IDE) to write, test, and iterate on data pipelines, spending your time on finding solutions rather than debugging code
- Write prompts that produce code you can actually trust. You will learn to verify agent output, catch mistakes, and maintain full control over your analysis
- Build predictive models with proper train/test splits and cross-validation, using SQL and Pandas for data processing
- Spot data leakage before it ruins a model in production. Use SHAP values to gain insights into the inner workings of complex models
- Turn messy unstructured text into usable features using LLMs. Transform documents, survey responses, or free-text fields into features enriching your model
- Complete an end-to-end project independently: from a raw dataset to a trained, validated model deployed as an interactive Streamlit application
THE CODEBRIDGE COMMUNITY
Your growth continues long after the course. As a graduate, you join an active community of professionals sharing real-world solutions and staying ahead of the rapidly evolving data landscape.
- Bi-Monthly Workshops: Focused sessions every two months covering a broad range of data topics, such as AI applications in healthcare, rapid web development, or new AI tooling.
- Monthly Meetups: Regular networking events to discuss ideas, troubleshoot project challenges, and build your professional circle.
- A Multi-Disciplinary Network: Connect with alumni and experts across medicine, academia, public policy, tech, and beyond.
PRICING OPTIONS
PRIVATE SECTOR
PUBLIC SECTOR
ACADEMIA
Would you like to become a partner and help us deliver cutting-edge education?
COURSE TEAM
Ján Dudek
Magna cum laude economics and econometrics graduate from Rice and Oxford. Improved the risk-equalization model at the Ministry of Health and implemented ML fraud detection algorithms in Slovak healthcare. Currently a senior data scientist specializing in the health insurance industry.
Imrich Berta
Applied mathematics graduate from University of Cambridge, experienced in machine learning models for disease prediction. Currently works as a consultant for government on cancer epidemiology and public health. Actively mentors analysts and organizes coding workshops for students.
Laura Johanesová
Laura is a bioinformatician and biomedical scientist currently at the University of Vienna. The skills she has in R, Linux and Python are crucial for her research in regeneration.
CONTACT US
Have any questions about the curriculum, prerequisites, or the agentic tools we will be using?
Drop us a line. Our team typically responds within 24 hours to help you determine if this advanced bootcamp is the right fit for your goals.