Built BhashaSub as a backend engineering project to learn production Python patterns. It's a FastAPI service that takes text input, calls the Gemini API to translate it into Indian languages, and returns the result through a browser UI. I focused on things like async communication, retry logic for rate-limited APIs, structured logging, and Pydantic validation — essentially the same patterns used in production AI pipelines. It works if someone pays for API keys 😶😶
- FastAPI REST API
- Async HTTP requests using HTTPX
- Environment-based configuration
- Structured logging
- Retry logic using Tenacity
- Pydantic request validation
- Swagger/OpenAPI documentation
- Pytest testing
- Python
- FastAPI
- HTTPX
- Pydantic
- Tenacity
- Pytest
pip install -r requirements.txt
uvicorn main:app --reload
Open: