This course is for engineers, researchers, and technically skilled practitioners who use convex optimization—and want to understand how solvers actually work. It’s ideal if:

• You’ve used gradient descent, CVXPY, or PyTorch optimizers

• You work with constraints, regularizers, or structured models

• You need to build or adapt algorithms—not just call a solver  

If black-box methods aren’t cutting it, this course is for you.

• How to decompose problems into modular components (e.g. proximals, gradients)
• How to build robust algorithms via combining operators
• How to translate theory into practical, reusable code

• Why operator properties like averagedness and cocoercivity guarantee stability

Clear explanations that connect to implementation
Concepts are explained with intuitive diagrams and connected to code features.

Modular, reusable methods
Build solvers from composable blocks like gradients, projections, and proximal operators.

Theory that’s actionable
Monotonicity, cocoercivity, and convergence are explained with practical structure in mind.

Beyond the basics
Go past gradient descent—learn Forward–Backward, Douglas–Rachford, ADMM, and more.

Linear Algebra – vectors, inner products, linear maps, matrix norms

Multivariable Calculus – limits, gradients, Hessians

Undergraduate Analysis –formal arguments, sequence convergence

Yes! The Operator’s Toolbox is designed with professionals in mind. When you enroll, you’ll receive:

• A completion certificate showing course hours, content covered, and instructor info

• A receipt/invoice suitable for employer reimbursement

• Optional documentation upon request (e.g., course outline or syllabus)

Learners can use their company’s CE, L&D, or professional development budgets to pay for this course. If you’re unsure, check with your manager or HR—they’re often happy to support career-aligned training like this.

The course is currently in development, but by pre-signing up, you'll be among the first to know when it's ready. Planned launch is September 2025.

No, this course focuses exclusively on continuous optimization.

The Operator’s Toolbox is built around first-order methods and monotone operator theory, which are designed for solving smooth, nonsmooth, and constrained problems in continuous spaces.

If you're looking for techniques like branch-and-bound, cutting planes, or integer programming, those are outside the scope of this course.