AI Development Lesson 10: Deploy an AI App

Building the model is 20% of the work. Deploying it so others can use it is the other 80%. This lesson covers the full stack from model to production.

FastAPI + AI Model

# pip install fastapi uvicorn pydantic
from fastapi import FastAPI
from pydantic import BaseModel
from groq import Groq

app = FastAPI(title="AI Code Reviewer", version="1.0")
client = Groq(api_key="...")

class ReviewRequest(BaseModel):
    code: str
    language: str = "python"

class ReviewResponse(BaseModel):
    bugs: list[str]
    suggestions: list[str]
    score: int

@app.post("/review", response_model=ReviewResponse)
async def review_code(req: ReviewRequest):
    response = client.chat.completions.create(
        model="llama-3.1-70b-versatile",
        messages=[
            {"role":"system","content":"Review code. Return JSON with bugs, suggestions, score(1-10)."},
            {"role":"user","content":f"Language: {req.language}\n\n{req.code}"}
        ],
        response_format={"type":"json_object"}
    )
    import json
    return json.loads(response.choices[0].message.content)

Streamlit for Quick UI

# pip install streamlit
import streamlit as st
from groq import Groq

st.title("AI Code Reviewer")
code = st.text_area("Paste your code:", height=200)

if st.button("Review"):
    with st.spinner("Analyzing..."):
        result = review_code(code)
    st.success(f"Score: {result['score']}/10")
    st.write("**Bugs Found:**")
    for bug in result['bugs']: st.error(bug)
    st.write("**Suggestions:**")
    for s in result['suggestions']: st.info(s)

# Run: streamlit run app.py