Big Tech Digest #11: 10 Unexpected Ways To Use GitHub Copilot, Building Smart Retry at Stripe, Stopping Fraudsters Through Penny Drop Verification at Uber and more!

2024-01-26 | Big Tech Digest #11

Happy Friday 👋!

This time around, I slightly evolved the format based on your feedback. From now on, every issue will start with the top 3 must-read articles followed by a number of optional reads. The goal is to make it easier and faster for you to skim through the issue.

Let me know what do you think!

I’m trying to increase my activity on X and share the articles ahead of time. Feel free to follow Big Tech Digest there, too!

One thing you could do to help me grow Big Tech Digest is to tell your connections about it. Any share is much appreciated 🙏!

Without further ado, let’s get started!

// 🏆 Top 3 must-reads

1. "10 unexpected ways to use GitHub Copilot"

by Kedasha Kerr ⸱ GitHub ⸱ 5 min read ⸱ 22 Jan

• Shows some cool Copilot features that allow to:

  • run terminal commands from Copilot Chat in VS Code

  • generate a pull request summaries with Copilot

  • create a commit messages with Copilot in VS Code

  • get help with terminal commands

  • explain repositories on GitHub

  • fix code inline

  • bulk close 1000+ GitHub issues with Copilot Chat

  • generate documentation for your code

  • get help with error messages and debug GitHub Actions workflow

• Definitely a must-read!

2. "How we built it: Smart Retries"

by Kiran Chandran ⸱ Stripe ⸱ 8 min read ⸱ 23 Jan

• Explores the use of machine learning algorithms to predict optimal time for retrying failed payments

• Describes the attributes used to train Smart Retries ML model

• Gives an overview of the heavy-duty retry models and their tradeoff between accuracy and speed

• Covers the use of multimodal data and embeddings for higher-precision predictions

3. “Stopping Uber Fraudsters Through Risk Challenges”

by Stephanie Yen ⸱ Uber ⸱ 7 min read ⸱ 25 Jan

• Discusses real-time fraud detection rules running on Mastermind

• Describes risk challenges and their role in catching bad actors

• Introduces penny drop verification as a method of fraud mitigation

• Gives an overview of the technical implementation of penny drop verification and presents the entire user flow

• Covers the process of sending authorization holds and amount verification

• Shares the continuous fine-tuning of user experience for the challenge

// 📬 Optional reads

a.k.a. The Best of the Rest!

“Warm up the relationship between Java and Kubernetes”

by Tony Demol ⸱ BlaBlaCar ⸱ 12 min read ⸱ 12 Jan

• Describes the migration of BlaBlaCar backend infrastructure to Kubernetes on Google Cloud Platform

• Discusses the delay in reaching peak performance for Java applications due to dynamically compiled byte code

• Shares the implementation of a warmup system leveraging Kubernetes native features to address cold JVM issues

• Covers the trade-offs and future alternatives for Java application warmup and performance optimization

"How Meta is advancing GenAI"

by Pascal Hartig ⸱ Meta ⸱ 1 min read ⸱ 11 Jan

• Covers the most interesting recent research papers on GenAI

• Introduces Meta’s latest GenAI innovations, including Audiobox, Emu, and Purple Llama

"Handling Online-Offline Discrepancy in Pinterest Ads Ranking System"

by many authors ⸱ Pinterest ⸱ 11 min read ⸱ 18 Jan

• Describes the challenges of translating offline model performance to online business metrics

• Discusses different scenarios, such as bug-free and buggy, in handling online-offline discrepancies

• Explores hypotheses for the observed online-offline discrepancy

• Shares methods to diagnose and resolve issues, including data issues and discrepancies between model training and serving logics

• Presents a case study on a real-life Pinterest issue and how it was resolved

"Canvas-based JavaScript Blur detection"

by Eremingeorgy ⸱ Revolut ⸱ 10 min read ⸱ 22 Jan

• Presents the challenge faced in developing a KYC module in the banking sector

• Discusses the absence of a suitable open-source JavaScript library for image blur detection

• Describes the limitations of existing solutions like OpenCV and plain JS approach

• Introduces a new lean and focused library designed for image blur detection in JavaScript/TypeScript

• Shares the approach and helper functions used in the proposed solution for image blur detection

"Uber: GC Tuning for Improved Presto Reliability"

by Cristian Velazquez, Vineeth Karayil Sekharan ⸱ Uber ⸱ 8 min read ⸱ 11 Jan

• Describes how Uber uses open-source Presto to query data sources

• Presents the memory fragmentation optimization activity carried out across all production clusters

• Introduces G1GC garbage collector used by Uber and its characteristics

• Explores the process of tuning JDK 11 for G1GC

• Shares the impact of GC tuning on Presto clusters at Uber

"Palette Meta Store Journey"

by many authors ⸱ Uber ⸱ 7 min read ⸱ 18 Jan

• Describes the challenges of querying Uber's data stores for feature data

• Shares the Uber Michelangelo feature store (Palette) as a solution for generating and using features in machine learning projects

• Explores the Palette Metastore as a metadata store for feature management infrastructure

• Goes through the major incident in 2021 due to inadequate schema validation on Palette Metadata

• Presents the re-architecture of the offline and online metadata update system, resulting in reduced deployment and customer onboarding time

"How to get in the flow while coding (and why it’s important)"

by Gwen Davis ⸱ GitHub ⸱ 6 min read ⸱ 22 Jan

• Shares quick tips for developers to achieve and maintain flow state, such as optimizing their environment and mapping out their work

That’s it for today! I hope you enjoyed this issue. Let me know in the comments what have you learned 💡!

Thanks for reading Big Tech Digest. Please don’t forget to 🗣️ spread the word and see you in two weeks!

---

Delivered bi-weekly to your inbox, Big Tech Digest brings you a collection of links to the latest engineering blog posts from +300 Big Tech companies and startups like Airbnb, Uber, Netflix or Meta. Aimed at Software Engineers and AI/ML folks at any level, Big Tech Digest is focused on interesting solutions to engineering problems that tech companies come across. No marketing or non-tech stuff.

Subscribe now to receive a new issue directly to your inbox every two weeks!