File and data processing
AUTOMATION & DATA RUNTIMEPythonAutomation, data processing & specialist services
DEVNU / TECHNOLOGY / PYTHON
We choose Python when its ecosystem is part of the solution, not merely because the syntax is concise.
Python has mature libraries for data processing, automation, computer vision and scientific work. In a multi-service architecture it can sit beside a TypeScript backend and own a specific workload.
What does Python own in a project?
At DEVNU, Python is more often considered for workers, file processing, automation and specialist services. A primary public API moves to Python when the ecosystem, team or workload genuinely supports that decision.
Where the technology should solve a specific problem.
Automation and scheduled jobs
Computer vision and OpenCV
ETL and data cleanup
Scientific-tool integrations
Specialist services behind queues or APIs
When we reach for it
Data processing, lightweight ETL, lawful controlled scraping, image processing, operational scripts and integrations with scientific tooling are natural fits.
When we leave it out
If the whole product is TypeScript and a new service merely duplicates a few simple endpoints in the same domain, a second runtime increases deployment, observability and skill-set cost without a clear benefit.
Using the tool is easy. Operating it well is the real work.
- 01Virtual environments and dependency pinning
- 02Type hints at important boundaries
- 03Worker isolation for heavy jobs
- 04Bounded timeouts and retries for integrations
- 05Structured logging
- 06Explicit input/output contracts between Python and other services
The decisions matter more than the stack badge.
Python or Node.js?
Python often wins for scientific and data ecosystems; Node.js often reduces friction for TypeScript APIs and web-centric integrations.
Script or service?
When work needs scheduling, retries, state or multiple consumers, a one-file script quickly becomes hidden infrastructure and deserves a clear service boundary.
Synchronous call or queue?
Slow or failure-prone processing is separated from the user request so a specialist service timeout does not block the entire experience.
Projects where Python is part of the recorded stack.
Telegram bot, Mini App & order automation
PasarGuard — Telegram commerce & operations
A personal product that brings ordering, wallet balance, service delivery and reseller operations into one Telegram bot and Mini App flow.
- Telegram
- React
- TypeScript
- FastAPI
- PostgreSQL
- Redis
- Docker
- Nginx
07 / RELATED SERVICES
Technology only matters inside a useful outcome.
Common questions about Python
01Does DEVNU use FastAPI?+
FastAPI is a candidate for typed lightweight Python services, but framework choice follows the workload rather than preceding it.
02Can Python and NestJS coexist?+
Yes. A Python service can own specialist processing and communicate with a NestJS backend through a queue or API.
09 / NEXT STEP
Start with the problem. Choose the stack second.
If the project needs this technology, the reason should be visible in architecture, performance, user experience or maintenance cost.
Describe the project