Model training
Models trained on your data, with a record of what they learned.
AI Development
TensorFlow development services beside our Python work
We have no published TensorFlow build. Our back ends run on Python 3.13, and we say plainly that none of them trains a model.
0Published builds on TensorFlow
2.21Latest stable release
3.9Python version TensorFlow 2.21 dropped
LiteRTWhere tf.lite is moving
What we build
What we would take on with TensorFlow
Serving behind an API
A FastAPI service in front of a model, as on our Python builds.
Upgrades
Moves onto 2.21 and a supported Python version.
On-device inference
Models on a phone through LiteRT, the successor to tf.lite.
Architecture
Where TensorFlow would sit
A model in the slot our AI product fills
Clients
React 19 appAccount briefs and grades
API
FastAPIPython 3.13
Model
TensorFlow modelWhere it would sit
Claude-compatible APISales OS, today
Illustration. A dashed outline marks where the technology would sit. We have no published build on it.
Edge
Mobile inference moves to LiteRT
From tf.lite to LiteRT
Release notes, 2.19 and 2.20
tf.lite
- tf.lite.Interpreter was to be deleted in TF 2.20
- tf.lite will be deprecated in favour of LiteRT
Replacement
LiteRT
- ai_edge_litert.interpreter replaces the old interpreter
- The LiteRT repository holds the code
Source: TensorFlow releases on GitHub, checked 4 November 2026. TensorFlow 2.21 also removed Python 3.9 support.
Builds
The closest builds we can show you
AI product, in-house
Sales OS
Python 3.13 and FastAPI with an AI engine that writes a graded brief through a Claude-compatible API.
1 creditOpens all six tabs of an account brief
Logistics, fleet operations
TruckFleet Labs
A Python 3.13 back end on FastAPI with telematics ingest and PostGIS.
38Drivers on the app by week three
Fit check
When TensorFlow is the right call
Pick TensorFlow when
- Your team already works in TensorFlow and Keras
- Models must run on a phone or an edge device
- You are on a release older than 2.21 and must upgrade
Look elsewhere when
- You need a published TensorFlow build to inspect, which we do not have
- A model API already answers the question, as in Sales OS
- You have no labelled data yet
Explore
Same job, other tools
Pairs with
Where we used it
Stack
What TensorFlow usually sits beside
Python
PyTorch
Docker
Kubernetes
OpenCV
scikit-learnFAQs
What buyers ask first
Which TensorFlow version is current?
TensorFlow 2.21 is the latest stable release on GitHub, and 2.22.0-rc0 is a pre-release. Version 2.21 removed Python 3.9 support.
Do you have a published TensorFlow build?
No. We say so on the first call and show the nearest work, which is a Python 3.13 back end and an AI product that calls a model API.
What happens to tf.lite?
The release notes say tf.lite is being deprecated in favour of LiteRT, with ai_edge_litert.interpreter replacing the old interpreter.
Who owns the code and the cloud accounts?
You do, from the first commit. Repositories and cloud accounts stay in your name.
Tell us what your model has to predict
Thirty minutes on your data, your labels and your TensorFlow version. No sales pitch.