AI Engineer (LLM & Generative AI)
Practice building LLM apps: prompts, embeddings, RAG, tool use and agents, evaluation, guardrails, fine-tuning and cost and latency trade-offs.
- Questions
- 100
- Interview Qs
- 30
- Judged problems
- 36
- Skills
- 9
About the role
An AI Engineer builds products on top of large language models and other generative models. Instead of training models from scratch, you combine pretrained models with prompts, retrieval, tools and application code to solve real problems such as support assistants, document search, coding helpers and workflow automation.
Day to day you design prompts and system instructions, build retrieval-augmented generation (RAG) pipelines with embeddings and vector search, connect models to tools and APIs, and write evaluation sets so you know whether a change made things better or worse. You also handle the hard production problems: hallucinations, prompt injection, data privacy, rate limits, latency and cost per request.
Good AI engineers are strong software engineers who understand how transformers, tokens and sampling work, and who treat model output as unreliable input that needs checking. They know when prompting is enough, when RAG helps, and when fine-tuning a smaller model is worth it.
Practise by level
Open the full question bank (100 questions) →Mock interview
Screening round
4 questions · 4 min each · pass 50%
mediumTechnical round
5 questions · 4 min each · pass 60%
hardFinal round
6 questions · 3 min each · pass 70%
Coding problems for ai engineer (llm & generative ai)s
- Best K-Day Sales Windoweasy
- Binary Classifier Metricseasy
- Cumulative Totalseasy
- Min-Max Feature Scalingeasy
- Most Common Allowed Wordeasy
- Sensor Moving Averageeasy
- Best Match by Cosine Similaritymedium
- Count Subarrays With a Given Summedium
- K Closest Feature Vectorsmedium
- Longest Shared Subsequencemedium
- Longest Span With K Distinct Productsmedium
- Most Frequent Valuesmedium
Skills in this track
Learning roadmap
- 1
Programming and API basics
Be solid in Python or TypeScript, HTTP APIs, async code and JSON. Call an LLM API, stream responses and handle errors and rate limits.
- 2
How LLMs work
Learn tokens, context windows, the transformer and attention at a high level, sampling (temperature, top-p) and why models hallucinate.
- 3
Embeddings and RAG
Build a RAG pipeline: chunking, embeddings, vector databases, hybrid search, reranking and citing sources. Measure retrieval quality.
- 4
Tools, agents and structured output
Use function or tool calling, JSON schemas and validation. Build small agents with clear step limits, and learn protocols for connecting tools.
- 5
Evaluation, safety and production
Create eval sets, use LLM-as-judge carefully, add guardrails against prompt injection, and optimise cost and latency with caching, smaller models and batching.
- 6
Fine-tuning when needed
Learn when fine-tuning beats prompting, and how parameter-efficient methods like LoRA work. Prepare clean training data and compare against a prompted baseline.
Top AI Engineer (LLM & Generative AI) interview questions
- Explain retrieval-augmented generation (RAG) to a new team member. Why do we use it?
- What are embeddings, and what can you build with them besides RAG?
- Why do LLMs hallucinate, and what simple steps reduce it in an application?
- How do tokens affect the cost and speed of an LLM application?
- What do temperature and top-p control, and how would you set them for different tasks?
- What is an LLM agent, and how is it different from a single LLM call?