Technology

TensorFlow

TensorFlow is Google's ML framework, strongest for production-grade model serving (TensorFlow Serving), mobile inference (TFLite), and large-scale training pipelines. TantraDev uses TensorFlow when the model leaves the notebook — when something has to be quantised for an Android phone, served behind a latency SLO, or trained on a multi-GPU cluster. For experimentation we are equally at home in PyTorch.

Where we use it

In production at TantraDev.

EdTech learning platform · adaptive difficulty

Related terms

Concepts that travel with this one.

Architecture rarely lives in isolation — these are the terms that come up in the same conversation.

TECHNOLOGY

Python

Python is TantraDev's default for data pipelines, ML inference, and back-office tooling. The library surface (Pandas, NumPy, scikit-learn, FastAPI, PyTorch) is wide enough that 'we'll prototype in Python' rarely turns into 'we have to rewrite it' — and where it does, the contract layer is small and explicit. We pair Python with strict type hints and Pydantic on the API edge.

Canonical references

ARCHITECTURE AUDIT

Building a system where TensorFlow is the load-bearing decision?

30 minutes on the phone, one page in your inbox — what to build, what to skip, what it will cost. You keep the audit even if we are not the right fit.

Book the audit