Every January, my inbox fills with the same kind of message. A financial analyst who has spent a decade in spreadsheets. A marketing manager who keeps getting asked for dashboards she cannot build. A healthcare operations lead who watches data teams present findings she knows she could produce. Different industries, same sentence underneath: I want to move into data, and I do not know where to start.
This roadmap is the answer I give them. It mirrors the structure of our 12-week Gen AI Data Science Bootcamp, because that sequence has worked for hundreds of career changers. You can follow it in a classroom or adapt it to your own pace. What matters is the order.
First, a word about your starting point
Career changers usually apologize for their background in the first conversation. Stop doing that. A finance professional already thinks in ratios, variance, and forecasts. A marketer already lives in funnels and segments. A healthcare operations person understands messy real-world data better than most computer science graduates. Your domain is an asset that a 22-year-old cannot buy. The roadmap below adds the technical layer on top of it.
The fear I hear most is about coding, so let me address it directly. You do not need a math degree or a childhood spent programming. You need basic comfort with numbers, which your current job almost certainly gave you, and a willingness to practice a few hours a week. Python is the most forgiving first language in wide use, and modern AI assistants make the early weeks gentler than they were five years ago. Every cohort I have taught included people who had never written a line of code. They finished at the same rate as the engineers.
The roadmap, phase by phase
Weeks 1-2: Foundations and Python
Start with the data science life cycle: how a business question becomes a dataset, a model, and a decision. Learn Python through Google Colab so nothing needs installing, and get comfortable with pandas for loading, cleaning, and exploring data. Build your first Tableau visualizations in parallel. By the end of week two you should be able to take a raw CSV and produce a short, honest summary of what is in it.
Weeks 3-4: Machine learning that earns trust
Regression first, then classification. Linear and logistic regression, decision trees, random forests, and the evaluation habits that separate real work from lucky guesses: train-test splits, cross-validation, precision and recall, and reading a confusion matrix without flinching. Apply each method to a business problem in your own industry. A churn model built by a marketer lands differently in interviews than a tutorial dataset ever will.
Week 5: NLP and text data
Most business data is text: reviews, tickets, notes, surveys. Learn tokenization, sentiment analysis, and how large language models can summarize and classify text at scale. This week tends to be the turning point for career changers, because text projects from their old jobs suddenly look like portfolio pieces.
Weeks 6-9: Feature engineering and agentic AI
Feature engineering is where models get good, and it rewards domain knowledge heavily. Then the agentic block: building AI agents with LangChain and AutoGen that pull data, run analysis, and draft reports under human supervision, plus the governance practices that keep those systems safe in a real company. During this stretch our students also complete the IBM Gen AI certificate, so the phase ends with a credential from a name every hiring manager recognizes.
Weeks 10-11: Resume and interview preparation
Translate your experience instead of hiding it. Your old job becomes the domain story; your projects become the proof. Practice the project walkthrough until you can defend every choice calmly, and prepare for the GenAI questions that now appear in nearly every data interview.
Week 12: A real project, delivered
Finish with end-to-end delivery. Our bootcamp closes with a freelance-style project, and our graduates often continue into the 6-week AI Practicum, a mentor-guided engagement with a real client. A delivered project changes interview conversations. You stop describing what you studied and start describing what you shipped.
What the timeline really demands
Our live classes meet Tuesdays 7-9 PM CT and Saturdays 2-4 PM CT, designed for people with day jobs and families. Expect several additional hours of practice each week. Twelve weeks of that rhythm is demanding and completely doable; our students hold a 70% job placement rate within six months of graduating, and nearly 1,000 people have completed the program. The students who struggle are the ones waiting to feel ready. The ones who finish are the ones who start on a date and protect the schedule. You will not do it alone, either. Every student gets 1:1 mentoring and a small cohort moving through the same weeks, and that accountability is the piece self-paced courses can never give you: people who notice when you fall behind and pull you back in before a missed week becomes a missed month.
Starting in January
If this roadmap sounds like your January, our next cohort is enrolling now. January enrollment is open with New Year Savings pricing at $1,495, against the regular $1,795. Come with your domain experience and your questions. We will handle the sequence, the mentoring, and the accountability.