Summarize & Sentiment API

Production-style FastAPI service that exposes LLM-powered text summarization and sentiment analysis endpoints using the OpenAI Responses API.

Built as an AI engineering portfolio project to demonstrate clean API design, schema validation, resilient model-output parsing, and deployment-ready structure.

Why This Project

This repository showcases practical applied-AI backend skills recruiters and hiring managers look for:

Designing typed, testable API contracts with Pydantic
Integrating LLMs into backend services (not just notebooks)
Hardening model outputs with parsing/validation logic
Implementing structured JSON logging (structlog) for observability
Organizing code into routes, services, and schemas for maintainability

Features

Health endpoint for service uptime checks
Summarization endpoint with length control
Sentiment endpoint returning structured JSON (sentiment, confidence, explanation)
Strict validation and normalization of model sentiment labels
Environment-driven config (OPENAI_API_KEY, optional OPENAI_MODEL)
Render-friendly entrypoint via main.py (uvicorn main:app)

Tech Stack

Python
FastAPI
OpenAI Python SDK (responses.create)
Pydantic v2
Structlog
Uvicorn
python-dotenv

Project Structure

.
├── app/
│   ├── main.py                 # FastAPI app factory, logging setup
│   ├── routes/
│   │   ├── health.py           # GET /health
│   │   ├── summarize.py        # POST /summarize
│   │   └── sentiment.py        # POST /analyze-sentiment
│   ├── schemas/
│   │   └── models.py           # Request/response models
│   └── services/
│       ├── summarize.py        # LLM summarization logic
│       └── sentiment.py        # LLM sentiment + robust JSON parsing
├── main.py                     # host entrypoint (uvicorn main:app)
└── requirements.txt

API Endpoints

`GET /health`

Returns status and timestamp.

`POST /summarize`

Generates a concise summary.

Request body

{
  "text": "Long source text...",
  "max_length": 80
}

Response body

{
  "summary": "Short summary text..."
}

`POST /analyze-sentiment`

Analyzes sentiment and returns typed output.

Request body

{
  "text": "I loved how smooth this release felt!"
}

Response body

{
  "sentiment": "positive",
  "confidence": 0.93,
  "explanation": "The text expresses clear satisfaction and positive emotion."
}

Local Setup

Clone and enter project

git clone https://github.com/zainabahmed4626-lab/AIEngineeringW1V2.git
cd AIEngineeringW1V2

Create virtual environment

python -m venv .venv
# Windows PowerShell
.venv\Scripts\Activate.ps1

Install dependencies

pip install -r requirements.txt

Configure environment variables

OPENAI_API_KEY=your_openai_key
# Optional
OPENAI_MODEL=gpt-4.1-mini

Run API

uvicorn app.main:app --reload

Alternative host-style run:

uvicorn main:app --reload

Open interactive docs

Swagger UI: http://127.0.0.1:8000/docs

Example cURL Commands

curl -X GET "http://127.0.0.1:8000/health"

curl -X POST "http://127.0.0.1:8000/summarize" \
  -H "Content-Type: application/json" \
  -d '{"text":"FastAPI makes building APIs fast and maintainable.","max_length":20}'

curl -X POST "http://127.0.0.1:8000/analyze-sentiment" \
  -H "Content-Type: application/json" \
  -d '{"text":"The onboarding flow is confusing and frustrating."}'

Engineering Notes

Sentiment service includes normalization for common label variants (pos, mixed, etc.) before mapping to strict enum values.
Services fail fast when required env vars are missing.
Route layer catches exceptions and returns controlled HTTP 500 errors while logging structured failure context.

Recruiter Snapshot

This project demonstrates readiness for roles involving:

AI/LLM backend engineering
API productization of GenAI features
Reliable model-in-the-loop service development
Deployable Python service architecture

If you are reviewing this repository for a role, I can also provide a short architecture walkthrough and trade-off discussion (latency, cost, and model reliability choices).