Skip to content

AI/ML · Career guide

AI Engineer

Build reliable products on top of machine learning and large language models.

Role overview

AI engineers integrate models into real software: retrieval-augmented generation, agents, evaluation, guardrails and cost/latency trade-offs. It's software engineering first, with enough ML understanding to know when a model is — and isn't — the right tool.

Typical responsibilities

  • Design AI features end to end
  • Build retrieval pipelines
  • Write evaluations and monitor quality
  • Control cost, latency and safety
  • Ship and iterate in production

Beginner roadmap

  1. 1

    Python

    Syntax, packaging, virtual environments, typing.

  2. 2

    Git

    Branching, pull requests, reviews.

  3. 3

    APIs

    HTTP, REST, auth, rate limits, streaming responses.

  4. 4

    SQL

    Querying, joins, indexes, Postgres.

  5. 5

    Machine learning basics

    Supervised learning, evaluation metrics, overfitting.

  6. 6

    LLMs

    Tokens, context windows, prompting, structured outputs.

  7. 7

    Retrieval (RAG)

    Chunking, embeddings, vector search, citations.

  8. 8

    Agents & tools

    Tool calling, planning, MCP concepts, guardrails.

  9. 9

    Cloud

    Deploying services, secrets, queues, observability.

  10. 10

    Production AI

    Evals, monitoring, cost control, safety, fallbacks.

  11. 11

    Portfolio

    Two deployed AI apps with evaluation write-ups.

  12. 12

    Interview preparation

    System design for AI features, coding, trade-off discussions.

Advanced directions

Fine-tuningModel evaluation at scaleMulti-agent systemsOn-device inferenceAI security (prompt injection defence)

Project ideas

  • beginner

    Study-notes Q&A bot

    Answer questions over your own notes with citations.

  • intermediate

    Support agent with tools

    Agent that looks up orders via tools, with an evaluation set.

  • advanced

    Evaluated RAG service

    Hybrid search, reranking, regression evals and cost dashboards.

Interview topics

  • RAG design
  • Embeddings
  • Evaluating LLM output
  • Prompt injection
  • Latency & cost trade-offs
  • Python coding
  • Designing an AI feature end to end

Resume guidance

  • Show how you evaluated quality, not just that you called an API.
  • Mention real constraints you handled: latency, cost, safety.