AI Engineering
Build products on top of large AI models.
AI engineers build applications using LLMs and foundation models — agents, copilots and AI features.
What the work looks like
Designing prompts and retrieval systems, evaluating outputs, building AI features into products.
Key skills
- Python/TypeScript
- LLM APIs
- Retrieval (RAG)
- Evaluation
- Product sense
Common tools
OpenAI/Groq APIsLangChain-style librariesVector databasesNext.jsPython
Pros
- Newest and fastest-moving field
- Low barrier to building
- Great for startups
Cons
- Tools change monthly
- Hype vs reality
- Career ladder still forming
A 12-month starter roadmap
Months 1–3
- • Python + APIs
- • Basic ML concepts
- • Build a chatbot
- • Git
Months 4–6
- • RAG systems
- • Embeddings & vector DBs
- • Evaluation techniques
- • Ship an AI app
Months 7–12
- • Agents & tool use
- • Fine-tuning basics
- • Production concerns (cost, latency)
- • Portfolio of 3 AI products
Reality check
Students like: Building something magical in a weekend.
Students dislike: Unpredictable model behaviour.
Hardest part: Making AI reliable and measurable.
Common myth: Prompting alone isn't a career — engineering skill is still the core.