Foundations
Python for Data & ML · Math for ML
Free practical PDF guide
The AI/ML roadmap I wish I had.
Find your starting point, see the skills current AI engineering roles keep asking for, and follow a core curriculum with concrete readiness checks.
It also maps the applied FersmithAI labs, gives you 5, 15, 30, and 60-minute study options, and ends with a five-session worksheet.
Free practical PDF guide
Send the complete roadmap to your inbox.
The 2026–27 AI engineer roadmap includes a starting-point audit, the core curriculum, applied labs, readiness checks, and a five-session worksheet.
Free. It arrives by email as a PDF. Unsubscribe anytime.
Inside the guide
FERSMITHAI
2026–27
AI/ML
Roadmap
A preview of the core path. The guide also shows where applied practice, evals, coding agents, and production work enter.
Python Essentials for Data
Feature Engineering
Transformers & Attention
Building AI Products
Start at your first unproven gap. Skip knowledge you can already demonstrate.
The path at a glance
The sequence has a job.
The roadmap starts with the tools every later module assumes. It then moves through modeling and modern architectures before ending with the work required to ship and maintain an AI product.
Data and models
Data Engineering & SQL · Machine Learning
Modern architectures
Deep Learning · LLMs & Generative AI
Shipping
MLOps & AI Engineering
Cross-checked against current roles
The same capabilities keep returning.
The title changes between companies. The work keeps circling back to software quality, data and model judgment, AI systems, evaluation, operations, and product decisions.
Software engineering
Python, clean code, tests, Git, APIs, Linux, and system design.
Data and ML judgment
SQL, data quality, statistics, model choice, experiment design, and error analysis.
LLM and agent systems
Grounding, retrieval, tool use, orchestration, context, and structured output.
Evaluation and reliability
Golden sets, graders, traces, regression checks, drift, guardrails, and human feedback.
Deployment and operations
Serving, containers, cloud, observability, cost, latency, retries, and safe releases.
Product and AI-assisted work
Clear specs, coding-agent direction, verification, and learning from users.
Every capability maps to a live FersmithAI module or Lab. The guide names the exact modules and gives you a small proof to produce before moving on.
Samples from the guide
Know what “ready” looks like.
Each module gives you a reason to learn it and a small piece of evidence you can produce before moving on.
Send me the complete guidePython Essentials for Data
Why: Write readable Python for data work.
Ready when: Load, clean, transform, and test a small dataset.
Feature Engineering
Why: Turn domain knowledge into measurable signal.
Ready when: Build a feature without leaking future information.
Transformers & Attention
Why: Read the architecture behind modern language models.
Ready when: Explain Q, K, V, masking, and the cost of long context.
Building AI Products
Why: Connect model behavior to user needs and risk.
Ready when: Ship a small loop with feedback, recovery, and a clear metric.
Who made this
Sinem Demirci
I finished a master’s in Computer Engineering–Information Technologies while flying full-time as cabin crew and raising my son.
I never had an engineering job to learn on, so I had to work out the order myself: what to read first, what could wait, and which skills kept appearing in role descriptions, engineering guides, and production failure reports.
This roadmap is that order. FersmithAI is the learning system I built around it.
Before it reaches your inbox
A few practical details.
What is included in the AI/ML roadmap?
The practical PDF includes a starting-point audit, a role map, the core curriculum with readiness checks, applied FersmithAI labs, time-based study options, a complete AI Agents module sample, and a five-session worksheet.
Who is this roadmap for?
It is for professionals, students, and career changers working toward AI engineering, ML engineering, data science, or data engineering. It also helps people already in the field organize gaps in their knowledge.
Does the roadmap create a FersmithAI account?
No. Entering your email sends the PDF to your inbox. You can read the complete AI Agents & Tool Use module without an account or card.
Is the roadmap free?
Yes. The PDF is free and arrives by email.
The complete roadmap is free
Put the full learning order in your inbox.
Send me the PDFPractical PDF guide · Delivered by email · No account required