AB Satyaprakash’s Post

Maturing is realising people are different, not abstract good or bad. In fact, right and wrong are almost always relative. In the realm of real-life interactions, it's common to encounter individuals whose lifestyles, achievements, thought processes, or work methods we detest. Such aversions can stem from differences in perspective, feelings of envy, perceived intellectual superiority, or myriad other reasons. What's intriguing, however, is that even amidst our fervent disdain for them, these individuals typically view their actions as morally upright, often considering them the only righteous path. Consider applying this perspective to your interactions on LinkedIn and share your thoughts in the comments below!

The number of people obsessed with IITs because they couldn’t get into one in their teens is crazy.

Software Engineering Life Tips - #01 Unless you are presenting something and absolutely need to be near your laptop during your daily standups, sync ups, or other meetings - Just put on your AirPods and walk around a little while you engage in discussions. The amount of desk time that most of us have is alarming. Especially for us younger folks while it might not seem like much now, it’ll eventually catch-up with age.

Saw a man French kissing a water fountain faucet at Bengaluru airport today. What’s wrong with people’s hygiene? 😶 PS. I was waiting in queue after him. Then I just decided to leave. Might as well have kissed him than drink water from the fountain. #bengaluru #airport #weird #experiences

Lazy is the new fast: How Lazy Imports and Cinder accelerate machine learning at Meta Meta has achieved a significant milestone in enhancing model training efficiency by adopting Lazy Imports and the Python Cinder runtime. This adoption has resulted in up to 40 percent improvements in time to first batch (TTFB) and a 20 percent reduction in Jupyter kernel startup times. These enhancements have accelerated experimentation capabilities and improved the overall ML developer experience (DevX). TTFB is critical in ML development, as it determines the speed at which an ML model transitions from idle to learning, impacting productivity and experimentation. Lazy Imports, coupled with Cinder, offer a comprehensive strategy to address the TTFB challenge. Unlike previous approaches, Lazy Imports with Cinder defers all imports until necessary, simplifying the startup process and reducing the need for meticulous import selection. This optimization has led to reduced runtime imports and streamlined development workflows. Meta's adoption of Lazy Imports has resulted in significant TTFB wins, enabling ML developers to reach the model training phase swiftly. However, the adoption of Lazy Imports also posed challenges. Compatibility issues with existing libraries such as PyTorch and NumPy required meticulous troubleshooting and adjustments. Balancing performance optimization with code dependability was crucial, as semantic changes introduced by Lazy Imports could affect code intuitiveness. Moreover, adopting new paradigms like Lazy Imports necessitated overcoming a learning curve for the development team. Looking ahead, Meta aims to streamline developer onboarding by investing in educational resources and onboarding materials. Enhancements in tooling are also underway to simplify debugging and troubleshooting processes associated with deferred imports. Additionally, Meta seeks to foster community collaboration to share insights, best practices, and address challenges collectively. Building a robust community around Lazy Imports is crucial for supporting adaptable paradigms and patterns aligned with this approach. #meta #engineering #ai #lazy #technology #innovation #development

Ha ha. This is just so wholesome 😂

Techies, Someone comes to you for a referral with a weak resume. What do you do? 1. Ask the person to make a better resume to increase their chances of getting through and risk the fact that they might hate you or feel bad 🙁 2. Ignore the person's request since you have to give referrals to 20 more people 😐 3. Just refer them anyways, and move on knowing there's slim to 0 chances of their resume being shortlisted 🙂 Would love to know your thoughts on this!

Tip for NoBroker.com - Maybe have ads and contact leads in 2 separate WhatsApp chats? I don’t wanna read about “Diwali ke safai services” while I am trying to look frantically for owner’s contact details 💀😂 On that note, Happy Diwali 🪔 #diwali #nobroker #househunting #bangalore #feature #request