Photo by Austin Ramsey As an engineering manager, crafting an engineering strategy is crucial to keeping your team aligned, productive, and focused on shared goals. Drawing from my experience and insights from Will Larson’s book “An Elegant Puzzle,” here’s my approach to develope an engineering strategy. 1. Implement Systems Thinking Systems thinking is essential for an engineering strategy. It involves recognizing how various parts of your organization interact and influence one another....

Photo by Geranimo One might think that managing a software team and an R&D team is pretty much the same game—after all, both involve highly technical, innovative minds working to push boundaries. But, the core difference lies in the nature and expectations of their outputs. Leading a software team is often about predictability and reliability. It’s project-driven with clear, defined objectives: deadlines, budget constraints, and customer requirements dictate the pace....

The world of machine learning has seen rapid advancements in recent years, leading to the widespread adoption of ML models across various industries. One key development in this domain is distributed federated learning, a decentralized approach that facilitates the efficient training and application of ML models at scale. In this blog, we will discuss the importance of distributed federated learning for businesses and its potential to revolutionize the way industries leverage ML models....

Fine-tuning a pre-trained model for an image classification task on a domain-specific problem can significantly reduce the time and computational resources required for training a deep neural network from scratch. In this tutorial, I will walk through the steps of fine-tuning a pre-trained ResNet-18 model on a custom dataset using PyTorch. Step 1: Load the pre-trained model let’s start by loading the pre-trained ResNet-18 model from the PyTorch model zoo. The PyTorch model zoo provides a collection of pre-trained models that can be used for various computer vision tasks....

Photo by Nicolas Hoizey Modern software engineering is all about shipping fast and often It was about a month ago that I was introduced to a book called Accelerate and later discovered another study Accelerate State of DevOps Report. It is all natural to agile teams today but what all these studies demonstrate is that engineering teams that deliver more often are more likely to meet or exceed their organizational performance goals....