On-device models
Models converted to TensorFlow Lite for Android and iOS, with quantisation to keep them small and fast.
Languages and frameworks
TensorFlow development.
TensorFlow is a machine learning framework with a strong path from training to production, including mobile, browser and edge targets. We use it when deployment to constrained devices or an existing TensorFlow pipeline is part of the job.
A short call. Your first build, a fixed quote, and the timeline. No obligation.
What we build with TensorFlow
Where it fits.
Keras training pipelines
Image and sequence models with tf.data input pipelines and callbacks for checkpointing and early stopping.
Forecasting models
Time series models for demand and usage prediction, evaluated against simple baselines first.
Production serving
Models exposed through TensorFlow Serving or Vertex AI endpoints with versioning and monitoring.
When it is the right choice
Choose TensorFlow when
- →The model has to run on a phone, microcontroller or in the browser.
- →You already have TensorFlow models and pipelines in place.
- →You deploy on Google Cloud and want its managed serving.
Worth knowing
Honest trade-offs.
- →Most new research code is published for PyTorch, so porting can take effort.
- →Version changes between releases can break older code, so we pin and test.
Often paired with
KerasTensorFlow LiteVertex AIPythonDockerMLflow
Related services
Services that use it.
Proof
Work built by the same team.
Closeum
13+ connected modules from calls through commissions
13+connected modules
5role types
32completed transactions
40+leads managed
SIIMS
Orders, production, warehouse, and dispatch on the same records
4operational roles
37backend tests passed
17hosted workflow tests passed
Ledgerline
603 tests passed in a dated financial-control verification
SAR 760Kcash reconciled end to end
603tests passed
14control findings fixed
3finance sources in one ledger
Common questions
Before you start.
TensorFlow or PyTorch for a new project?
PyTorch is our default for new work. We choose TensorFlow when edge deployment or an existing codebase points that way.
Can a model really run on a phone?
Yes, if it is small enough. We convert and quantise it, then test accuracy and speed on real devices.
How do you check the model is any good?
We set a metric and a baseline before training, hold out data it never sees, and test on realistic inputs including awkward ones.
What do you build with TensorFlow?
Model training, On-device models, Computer vision, Forecasting, and related work. We choose TensorFlow when it is the right fit for the product, the team and the budget.
Can you work with our existing TensorFlow code?
Yes. We can review what you have, fix what is fragile, and extend it, or build alongside your team.
Keep exploring
More than TensorFlow.
Where we apply it
The Audit
Get my free AuditKnow what to build first.
Book a short call and show us one workflow. You leave with what we would build first, a fixed quote, and what should wait.