What you will learn
🔥 Enrollment Open — Applied Conformal Prediction
The most practical, zero-hype course on Conformal Prediction—designed to take you from curiosity to confident deployment.
If you’re tired of vague uncertainty and black-box trust issues, this is for you. Seats are capped to keep the cohort high-signal.
Built for real-world ML practitioners
Add CP to classification, regression, time series, CV, and NLP pipelines
Get distribution-free uncertainty with measurable coverage guarantees
Deliver reliable prediction sets/intervals and communicate risk clearly
Deploy CP in production with modern ML tooling
Instructor: Dr. Valeriy Manokhin
Trained under Prof. Vladimir Vovk, author of Practical Guide to "Applied Conformal Prediction" and "Applied Conformal Prediction: Reliable Uncertainty Quantification for Real-World Machine Learning in Python", creator of the Awesome Conformal Prediction repo, and a published ML researcher. Experience building and advising production AI systems across industry.
If your models drive real decisions, uncertainty is mission-critical. Join the next cohort and deploy CP with confidence.
What you’ll learn
Master Conformal Prediction and become the expert who delivers trustworthy, risk-aware ML in production.
Learn how to add CP to classification, regression, time series, CV, and NLP pipelines without breaking validity.
Produce prediction sets and intervals with finite-sample, distribution-free coverage guarantees.
Integrate CP into real ML systems.
Recognize efficiency–validity trade-offs, common pitfalls, and when CP assumptions are violated.
Become the expert who can defend uncertainty, challenge heuristics, and deliver risk-aware ML systems.
Explain coverage guarantees, assumptions, and trade-offs so stakeholders trust the uncertainty you ship.
Learn directly from Valeriy
Valeriy Manokhin
Conformal Prediction expert · PhD · CP book author
Who this course is for
Senior Data Science Leaders at Global Brands
Titles: VP, Director of Data & Analytics, Head of AI
Mid-Level Data Scientist in Tech, Finance, or Industry
Titles: Lead / Senior Data Scientist, Quantitative Researcher, ML Engineer
PhD Researcherc/ Academic in Data Science, Consultant
Titles: Professor, PhD Student, Assistant Professor, Researcher, Consultant
Prerequisites
High-school level math
Comfort with high-school level math, including basic algebra, probability, and statistics. No advanced math or calculus required.
🐍 Python
Basic proficiency in Python. You should be able to read and write simple code using tools like pandas, numpy, and scikit-learn.
No Conformal Prediction background - not needed
No prior experience with Conformal Prediction is required—everything you need is explained from the ground up.
What's included
Live sessions
Learn directly from Valeriy Manokhin in a real-time, interactive format.
♾️ Lifetime Access
Revisit all course materials, recordings, and resources anytime you need—no expiration, no gatekeeping. Use it as a long-term reference.
🌐 Community of Global Peers
Join a private community of forecasting professionals from top companies like Amazon, Apple, Google, Goldman Sachs, MorganStanley, Meta
📜 Certificate of Completion
Earn a professional certificate to showcase your forecasting expertise—perfect for sharing with your employer or featuring on LinkedIn.
🌐 Community of Global Peers
Join a private community of professionals from top companies like Amazon, Apple, Google, Goldman Sachs, Morgan Stanley. Walmart, Target.
Maven Guarantee
This course is backed by the Maven Guarantee. Students are eligible for a full refund up until the halfway point of the course.
Course syllabus
Week 1
Applied Conformal Prediction | Lesson 1
Applied Conformal Prediction | Lesson 2
Applied Conformal Prediction | Office Hours #1
Week 2
Applied Conformal Prediction | Lesson 3
Applied Conformal Prediction | Lesson 4
Applied Conformal Prediction | Office Hours #2
Special guests talk
Free resource
Unlock Accurate AI Forecasts with Superior Metrics
Why MAPE Fails and How to Choose Better Forecasting Metrics
Learn why MAPE misleads forecasters, with clear examples and practical steps to choose accurate metrics confidently.
Schedule
Live sessions
4 hrs / week
Projects
1 hr / week
Async content
1 hr / week
Testimonials
- Amazing course! I believe this course is quite unique, and you can't find a similar one anywhere else. Valeriy is one of the few instructors who combine both academic and industry expertise, and the discussions in the course were quite useful. The material is dense for the short course time, so I recommend getting the book beforehand
Mohammad Abdollah
PhD, Politecnico di Milano - I am demanding—both with myself and with the technologies and methods I choose to use. I've completed many specializations in the past, but looking back, I now see most of them as mere introductions, not directly applicable to real-world problems. Last year, I took the 'Conformal Prediction' course and found it to be of exceptional quality.
Stefan Degeye
Data Analysis Manager at Elia Group, became the winner of the first Kaggle Probabilistic Forecasting competition. - For too long, uncertainty in decision sciences was treated as a niche, often overlooked. Point predictions ruled, making trust in models hard—especially in high-stakes settings. The Conformal Prediction course changed that for me, offering a clear, practical path to statistically valid, trustworthy insights.
Chirag Ahuja
Senior Data Scientist with Visteon Coorporation - The Conformal Prediction Course deepened my understanding of classification. I learned to use methods like Venn-Abers and saw how conformal prediction tackles imbalanced data. It inspired my Python library, TinyCP, now used in my work. I’ve dropped SMOTE. For regression, guaranteed prediction intervals blew my mind. 10/10—highly recommend!
Lucas Leão
Lead Data Scientist - Valeriy’s course is a must for data scientists. It blends fundamentals and real-world use of Conformal Prediction to quantify and act on uncertainty. His clear teaching makes complex ideas accessible. At Target, it’s key to designing resilient supply chains—accuracy alone isn’t enough without understanding uncertainty.
Prajwal Sreenivas
Lead Data Scientist, Target
Frequently asked questions
$950
USD