Starting AI career — what projects should I build after learning Python + ML basics?

[ManuManohar]

Hi everyone,

I’m starting my journey toward becoming an AI engineer and currently learning Python and ML fundamentals.

I’d love advice from people working with LangChain or LLM apps:
What projects should I build next to move toward AI/LLM engineering?

Thank you!

[keenborder786]

• This guide is specifically for Applied Agentic Engineering with a focus on Langchain ecosystem.
  • Start by understanding the following core component: Agents - Docs by LangChain
    • Agents
    • Models
    • Tools
    • Messages
    • Short-Term Memory
    • Structured Output
  • Get a Basic Idea of Middewares: Overview - Docs by LangChain
  • Understand the Very Basic of Langgrah: Quickstart - Docs by LangChain
  • Once you have a clear Idea about Langchain and how Langchain v1 components are built on top of Langgraph (understand how they are linked), start building some stuff:
    • langgraph/examples at main · langchain-ai/langgraph · GitHub (You can take inspiration from these examples)
  • Also, learn about Langsmith: Tracing quickstart - Docs by LangChain and how you can integrate it in your workflow for better observability
  • Finally, to deploy your flow, I would recommend you learn:
    • https://docs.langchain.com/oss/python/langgraph/local-server (If you want to deploy your Agentic Flow using Langgraph template)
    • FASTAPI (Learn basics about FastAPI since that is the go-to framework for deploying most of ML based apps)

[Bitcot_Kaushal]

Hi @ManuManohar welcome!

Great to hear you’re starting your AI journey Since you’re already learning Python + ML basics, a good next step is to move toward LLM application building, not just model training.

Here’s a practical progression:

1. Build a Simple RAG App

• Load PDFs or text files

• Store embeddings in a vector DB

• Build a Q&A chatbot over your own data

This teaches retrieval, embeddings, and prompting.

---

2. Build a Tool-Using Agent

• Create tools (weather API, calculator, DB query)

• Use create_agent to let the model decide which tool to call

This teaches tool calling + reasoning.

---

3. Build a Multi-Step Workflow with LangGraph

• Example: booking assistant, research agent, or support bot

• Add memory, state, and conditional branching

This teaches orchestration (very valuable skill).

---

4. Deploy Something

• Add streaming

• Deploy on FastAPI

• Connect to a frontend

Real-world deployment experience is a big differentiator.

---

If you consistently build and ship small but complete apps, you’ll grow much faster than just doing notebooks.

You’re on the right path — keep building