# TensorFlow developer — alongside PyTorch, since 2024

> Model training with TensorFlow: where I have used the framework since 2024 and how it sits with PyTorch, CUDA and model serving.

- Highlights: 2 years
- Experience: Since 2024
- Updated: 2026-11-08
- Source: https://muhammetsafak.com/expertise/tensorflow/
- Language: en-US
- Author: Muhammet Şafak

---
## Me in the TensorFlow ecosystem

TensorFlow is one of the two frameworks in my model-training work. The other is [PyTorch](https://muhammetsafak.com/expertise/pytorch/); I see them not as rivals but as tools I choose by the job at hand.

Training is only one part of the system. Running on the GPU depends on CUDA, and putting the trained model behind an API depends on NVIDIA Dynamo-Triton; how these parts sit together is on the [AI systems engineering page](https://muhammetsafak.com/expertise/ai-systems/).

I have worked with TensorFlow since 2024. The page is as short as my claim.

## What is the TensorFlow Developer role?

TensorFlow is an open-source machine learning platform developed by Google: it provides tools to define, train and deploy models. A TensorFlow developer is the person who builds and trains a model with those tools.

## What does the TensorFlow Developer do?

I use TensorFlow in my model-training work, writing in Python and running training on the GPU with CUDA. PyTorch is part of the same set of work. To put trained models behind an API I use NVIDIA Dynamo-Triton.

## Technologies I pair it with

- TensorFlow: The deep learning framework I use in my model-training work.
- Python: The main language of my model-training, fine-tuning and data-preparation work.
- CUDA: The NVIDIA platform I run training and inference workloads on the GPU with.
- PyTorch: The deep learning framework I use for model training, fine-tuning and adaptation with LoRA/PEFT.
- NVIDIA Dynamo-Triton: The model serving server, formerly NVIDIA Triton Inference Server, that I use to put trained models behind an API.

## Frequently asked

### How long have you worked with TensorFlow?

Since 2024, which is 2 years.

### What does a TensorFlow developer do?

They define, train and evaluate a model in TensorFlow. In my work this is part of model training; data preparation and training run in Python.

### TensorFlow or PyTorch?

I use both in my model-training work and I make no claim that one is superior. On the [PyTorch page](https://muhammetsafak.com/expertise/pytorch/) I cover fine-tuning and adaptation with LoRA/PEFT; this page covers where TensorFlow sits.

### Do you use a GPU with TensorFlow?

Yes. I run training and inference workloads on the GPU with NVIDIA's CUDA platform; the details are on the [CUDA page](https://muhammetsafak.com/expertise/cuda/).

### How do you serve a model trained in TensorFlow?

With NVIDIA Dynamo-Triton, the serving server formerly called NVIDIA Triton Inference Server, which puts a trained model behind an API.

### Which language do you use TensorFlow with?

Python. It is the main language of my model-training, fine-tuning and data-preparation work; the [Python page](https://muhammetsafak.com/expertise/python/) covers the language itself.