Home / Roadmaps / AI & Machine Learning
AI & Machine Learning
From the maths of models to building with modern LLMs.
This path goes deeper than data science into how models actually work, plus the practical side of building applications on top of today's foundation models. It is demanding but among the highest-leverage skills you can have right now.
The route
Maths, Python & classical ML 8–12 weeks
You can't skip the fundamentals. Calculus intuition, linear algebra, and scikit-learn.
- Python
- Linear algebra & calculus
- scikit-learn
- Model evaluation
Deep learning 8–12 weeks
Neural networks, training dynamics, and one framework you know well.
- PyTorch
- Neural network training
- CNNs / RNNs
- Transfer learning
LLMs & building with AI 6–10 weeks
Prompting, retrieval, and shipping applications on top of foundation models.
- Prompt engineering
- RAG & embeddings
- APIs (OpenAI/Anthropic)
- Evaluation
MLOps & a real system 8–12 weeks
Deploy, monitor, and maintain a model in production — the part most courses skip.
- Model serving
- Monitoring & drift
- Docker & cloud
- Capstone project
Who it's for
- Data scientists going deeper
- Engineers building AI features
- Researchers and ambitious self-learners
Career outcomes
- ML Engineer
- AI Engineer
- Applied Scientist
Prefer guided learning?
If self-paced study keeps stalling, a mentored cohort can get you to job-ready faster. DTI runs structured programs in Albania across software, data, and security.
Explore DTI →