Wall Street Prep is best known for teaching bankers to build three-statement models in Excel, and this programme is its attempt to do the same thing for AI, with Columbia Business School Executive Education putting its name on the certificate. I expected a slide-deck course about AI strategy with a chatbot demo at the end. It is more substantial than that, and that is the main reason it rates as well as it does. The structure is eight modules over eight weeks.
Week one is the gentle overview of machine learning and generative AI, plus how to use chatbots in your own work (there is now an optional intro to Claude Cowork, though you need your own Claude subscription for it). Then it does something most finance AI courses avoid: it teaches Python. Module two covers Python basics for business, getting Excel and Python talking to each other, and calling APIs to automate workflows. Module three moves into pandas and Matplotlib on bigger datasets.
Modules four and five are classic predictive analytics, with linear and logistic regression, overfitting and validation, then decision trees and random forests applied to things like predicting loan defaults and early mortgage repayments. There is a hedge fund exercise on predicting stock returns and backtesting a strategy. Module six is simulation and optimisation, which is the most Columbia part of the course and the part I would least expect from a typical AI certificate. The last two modules come back to generative AI with embeddings, LLM-based classification, fraud monitoring on payment networks, retrieval augmented generation for financial analysis and building a basic agent.
That is a sensible syllabus. It is roughly what a good first-year MBA analytics elective covers, pointed at finance problems, with modern LLM material on top. The teaching team is the other real strength. Daniel Guetta, who directs the programme, is a former Palantir data scientist who co-wrote Python for MBAs, and the faculty list includes Ciamac Moallemi, Olivier Toubia and Stijn Van Nieuwerburgh.
Those are people who actually teach this material to Columbia students. The guest speaker list runs to around thirty names from Morgan Stanley, BlackRock, Ramp, Hebbia, Rogo, Perplexity and others. I would not put much weight on that. Recorded guest talks are pleasant, but they are marketing for the programme as much as teaching.
Now the problems. The first is price. $5,000 for eight weeks of mostly self-paced video with weekly office hours is a lot. Wall Street Prep's own comparison pages put it against programmes from MIT and Wharton at similar or higher prices, and against that peer group it looks fine.
Against the rest of the market, where Python for finance, scikit-learn and RAG tutorials are all available cheaply or free, it is expensive. What you are paying for is the structure, the deadlines, the office hours and the Columbia name. The second is pace. The marketing says no coding, Excel or statistics knowledge is required.
I believe that a motivated beginner can get through the Python modules. I am less convinced that someone who wrote their first line of code in week two will really understand embeddings and agent building by week eight, at ten hours a week. People with some prior Excel modelling or a little code will get far more from it. The third is the credential.
It is a Certificate of Participation from Columbia Business School Executive Education. That is a reasonable line on a CV and a decent conversation starter in finance, where the brand carries weight. It is not academic credit, and nobody should present it as a Columbia qualification. The refund policy is worth reading carefully too.
Tuition is non-refundable, but if you cannot finish you are automatically moved into the next cohort once at no cost. That is fairer than a flat no-refund rule, and the $550 tuition assistance available to anyone who applies is a nice touch. My 3.7 reflects a well-designed, properly taught programme that is priced for employer reimbursement rather than for individuals. If your firm will pay, I think it is one of the better uses of a training budget in this category, because you leave with Python skills and a set of worked finance projects, not just a vocabulary.
If you are paying out of your own pocket, the same skills are available far more cheaply, and the money is better spent on time to build things.