Custom TensorFlow Classification Model -- 4

The project centers on building a production-ready TensorFlow 2.x model that classifies tabular data delivered to us through an internal API. I have the API specifications and sample payloads ready; you will turn those streams into a clean training pipeline, engineer the right features, and iterate until the classifier meets our performance targets in real-world tests.

Scope of work
• Data pipeline – pull the API data, handle preprocessing, and produce TensorFlow-friendly datasets for train/val/test splits.
• Model development – design, train, and tune a deep learning architecture suitable for tabular inputs (e.g., wide & deep, Transformer, or other proven structures).
• Optimization – experiment with hyperparameters, regularization, and callbacks to achieve stable accuracy and minimal inference latency.
• Packaging – export the finished model as a SavedModel plus concise Python inference wrapper, ready to drop into our existing service.
• Integration – provide a small integration script/snippet that consumes live API data, calls the model, and returns the predicted class in the required JSON format.

Acceptance criteria
1. End-to-end notebook or script demonstrates ≥ X% validation accuracy (we will agree on the exact threshold upfront).
2. Inference time per record ≤ Y ms on our target hardware.
3. Codebase follows PEP 8, is reproducible with requirements.txt, and runs under Python 3.10.
4. Delivered artefacts: source code, trained model files, brief README covering setup, training, and deployment steps.

If you routinely work with TensorFlow tabular pipelines, understand data APIs, and can hand off clean, production-grade code, this assignment should feel straightforward. I will supply API keys, schema docs, and label definitions as soon as we start.