Scaling Grab's Data Lake: Our journey to Apache Iceberg adoption
As Grab's Data Lake grew to petabytes, the limitations of Hive Parquet became clear, from catalog latency and small files to manual partition management. This blog shares our journey adopting Apache Iceberg as our default table format: why we chose it, how we rolled it out at scale, the tooling we built along the way, and the lessons we learned.
More articles
Migrating Counter Service storage: Design choices and learnings
Scaling out Distroless adoption With AI
Palana (Part 2): Architecting isolation, identity, and auditability for AI agents
Palana (Part 1): Why Grab built a secure platform for autonomous AI Agents
From decentralized Docs-as-Code to a centralized repository: Evolving Grab's documentation strategy
The Hugo evolution: Engineering Grab's unified, one-click data ingestion platform with Apache Flink
Scaling developer experience: How we improved Android Studio in a large monorepo
Enhancing Flink deployment with shadow testing
Data Mesh at Grab (Part II): The foundational tools behind certification
Articles
Scaling Grab's Data Lake: Our journey to Apache Iceberg adoption
Migrating Counter Service storage: Design choices and learnings
Scaling out Distroless adoption With AI
Palana (Part 2): Architecting isolation, identity, and auditability for AI agents
Palana (Part 1): Why Grab built a secure platform for autonomous AI Agents
From decentralized Docs-as-Code to a centralized repository: Evolving Grab's documentation strategy
The Hugo evolution: Engineering Grab's unified, one-click data ingestion platform with Apache Flink
Scaling developer experience: How we improved Android Studio in a large monorepo
Enhancing Flink deployment with shadow testing
Data Mesh at Grab (Part II): The foundational tools behind certification
From firefighting to building: How AI agents restored our team’s core productivity
Enabling R8 optimization at scale with AI-assisted debugging
Reclaiming Terabytes: Optimizing Android image caching with TLRU
Cursor at Grab: Adoption and impact
Docker lazy loading at Grab: Accelerating container startup times
From deployment slop to production reality: How BriX bridges the gap with enterprise-grade AI infrastructure
Demystifying user journeys: Revolutionizing troubleshooting with auto tracking
How Grab is accelerating growth with real-time personalization using Customer Data Platform scenarios
A Decade of Defense: Celebrating Grab's 10th Year Bug Bounty Program
Real-time data quality monitoring: Kafka stream contracts with syntactic and semantic test
SpellVault’s evolution: Beyond LLM apps, towards the agentic future
Grab's Mac Cloud Exit supercharges macOS CI/CD
How we built a custom vision LLM to improve document processing at Grab
Machine-learning predictive autoscaling for Flink
Modernising Grab’s model serving platform with NVIDIA Triton Inference Server
Highly concurrent in-memory counter in GoLang
User foundation models for Grab
Powering Partner Gateway metrics with Apache Pinot
Taming the monorepo beast: Our journey to a leaner, faster GitLab repo
Data mesh at Grab part I: Building trust through certification
The evolution of Grab's machine learning feature store
Grab's service mesh evolution: From Consul to Istio
DispatchGym: Grab’s reinforcement learning research framework
Counter Service: How we rewrote it in Rust
The complete stream processing journey on FlinkSQL
Effortless enterprise authentication at Grab: Dex in action
From failure to success: The birth of GrabGPT, Grab’s internal ChatGPT
Streamlining RiskOps with the SOP agent framework
Introducing the SOP-driven LLM agent frameworks
Evaluating performance impact of removing Redis-cache from a Scylla-backed service
Facilitating Docs-as-Code implementation for users unfamiliar with Markdown
Improving Hugo stability and addressing oncall challenges through automation
Building a Spark observability product with StarRocks: Real-time and historical performance analysis
TechDocs at Grab: Cultivating a culture of quality documentation
Grab AI Gateway: Connecting Grabbers to multiple GenAI providers
Embracing passwordless authentication with Grab’s Passkey
Turbocharging GrabUnlimited with Temporal
How we seamlessly migrated high volume real-time streaming traffic from one service to another with zero data loss and duplication
Supercharging LLM application development with LLM-Kit
How we reduced initialisation time of Product Configuration Management SDK
Metasense V2: Enhancing, improving and productionisation of LLM powered data governance
How we reduced peak memory and CPU usage of the product configuration management SDK
LLM-assisted vector similarity search
Leveraging RAG-powered LLMs for analytical tasks
Evolution of Catwalk: Model serving platform at Grab
Enabling conversational data discovery with LLMs at Grab
Bringing Grab’s Live Activity to Android: Enhancing user experience through custom notifications
Unveiling the process: The creation of our powerful campaign builder
Chimera Sandbox: A scalable experimentation and development platform for Notebook services
How we improved translation experience with cost efficiency
LLM-powered data classification for data entities at scale
Profile-guided optimisation (PGO) on Grab services
How we evaluated the business impact of marketing campaigns
No version left behind: Our epic journey of GitLab upgrades
Ensuring data reliability and observability in risk systems
Grab Experiment Decision Engine - a Unified Toolkit for Experimentation
Iris - Turning observations into actionable insights for enhanced decision making
Android App Size at Scale with Project Bonsai
Enabling near real-time data analytics on the data lake
The journey of building a comprehensive attribution platform
Managing dynamic marketplace content at scale: Grab's approach to content moderation
Rethinking Stream Processing: Data Exploration
Kafka on Kubernetes: Reloaded for fault tolerance
Championing CyberSecurity: Grab's bug bounty programme in 2023
Sliding window rate limits in distributed systems
An elegant platform
Road localisation in GrabMaps
Graph modelling guidelines
Scaling marketing for merchants with targeted and intelligent promos
Stepping up marketing for advertisers: Scalable lookalike audience
Building hyperlocal GrabMaps
Streamlining Grab's Segmentation Platform with faster creation and lower latency
Unsupervised graph anomaly detection - Catching new fraudulent behaviours
Zero traffic cost for Kafka consumers
Go module proxy at Grab
PII masking for privacy-grade machine learning
Performance bottlenecks of Go application on Kubernetes with non-integer (floating) CPU allocation
How we improved our iOS CI infrastructure with observability tools
2.3x faster using the Go plugin to replace Lua virtual machine
Safer deployment of streaming applications
Message Center - Redesigning the messaging experience on the Grab superapp
Evolution of quality at Grab
How OVO determined the right technology stack for their web-based projects
Migrating from Role to Attribute-based Access Control
Securing GitOps pipelines
New zoom freezing feature for Geohash plugin
Graph service platform
Zero trust with Kafka
How KartaCam powers GrabMaps
Graph for fraud detection
Query expansion based on user behaviour
Using mobile sensor data to encourage safer driving
Automatic rule backtesting with large quantities of data
How we store and process millions of orders daily
How we automated FAQ responses at Grab
Graph Networks - 10X investigation with Graph Visualisations
How facial recognition technology keeps you safe
Graph concepts and applications
Automated Experiment Analysis - Making experimental analysis scalable
Search architecture revamp
Embracing a Docs-as-Code approach
Graph Networks - Striking fraud syndicates in the dark
How we reduced our CI YAML files from 1800 lines to 50 lines
How Kafka Connect helps move data seamlessly
Supporting large campaigns at scale
How telematics helps Grab to improve safety
Real-time data ingestion in Grab
Abacus - Issuing points for multiple sources
Exposing a Kafka Cluster via a VPC Endpoint Service
How Grab built a scalable, high-performance ad server
Biometric authentication - Why do we need it?
Using real-world patterns to improve matching in theory and practice
Designing products and services based on Jobs to be Done
Search indexing optimisation
Automating Multi-Armed Bandit testing during feature rollout
How We Cut GrabFood.com’s Page JavaScript Asset Sizes by 3x
Protecting Personal Data in Grab's Imagery
Processing ETL tasks with Ratchet
App Modularisation at Scale
Reshaping Chat Support for Our Users
Debugging High Latency Due to Context Leaks
Building a Hyper Self-Service, Distributed Tracing and Feedback System for Rule & Machine Learning (ML) Predictions
Our Journey to Continuous Delivery at Grab (Part 2)
How We Improved Agent Chat Efficiency with Machine Learning
How Grab Leveraged Performance Marketing Automation to Improve Conversion Rates by 30%
One Small Step Closer to Containerising Service Binaries
Customer Support Workforce Routing
Serving Driver-partners Data at Scale Using Mirror Cache
The GrabMart Journey
Trident - Real-time Event Processing at Scale
Pharos - Searching Nearby Drivers on Road Network at Scale
Reflecting on the Five Years of Bug Bounty at Grab
How Grab is Blazing Through the Superapp Bazel Migration
Democratising Fare Storage at Scale Using Event Sourcing
Keeping 170 Libraries Up to Date on a Large Scale Android App
Optimally Scaling Kafka Consumer Applications
Our Journey to Continuous Delivery at Grab (Part 1)
Uncovering the Truth Behind Lua and Redis Data Consistency
Securing and Managing Multi-cloud Presto Clusters with Grab’s DataGateway
Go Modules- A Guide for monorepos (Part 2)
The Journey of Deploying Apache Airflow at Grab
How We Built Our In-house Chat Platform for the Web
Go Modules- A Guide for monorepos (Part 1)
Does Southeast Asia Run on Coffee?
GrabChat Much? Talk Data to Me!
7 Fun Facts about Grab’s Driver-Partners in Singapore
Tackling UI Test Execution Time Imbalance for Xcode Parallel Testing
Returning 575 Terabytes of Storage Space to Our Users
Grab-Posisi - Southeast Asia’s First Comprehensive GPS Trajectory Dataset
How We Prevented App Performance Degradation from Sudden Ride Demand Spikes
Plumbing At Scale
Journey to a Faster Everyday Superapp Where Every Millisecond Counts
Marionette - Enabling E2E User-scenario Simulation
How We Implemented Domain-Driven Development in Golang
Driving Southeast Asia Forward Through People-Focused Design
Griffin, an Anti-fraud Risk Rule Engine Making Billions of Predictions Daily
Using Grab’s Trust Counter Service to Detect Fraud Successfully
Being a Principal Engineer at Grab
Data First, SLA Always
Save Your Place with Grab!
No More Forgetting to Input ERP Charges - Hello Automated ERP!
How We Built a Logging Stack at Grab
Making Grab’s Everyday App Super
Catwalk: Serving Machine Learning Models at Scale
React Native in GrabPay
Connecting the Invisibles to Design Seamless Experiences
Tourists on GrabChat!
Bubble Tea Craze on GrabFood!
Why You Should Organise an Immersion Trip for Your Next Project
Preventing Pipeline Calls from Crashing Redis Clusters
Guiding You Door-to-Door via Our Superapp!
Loki, a Dynamic Mock Server for HTTP/TCP Testing
How We Harnessed the Wisdom of Crowds to Improve Restaurant Location Accuracy
Designing Resilient Systems Beyond Retries (Part 3): Architecture Patterns and Chaos Engineering
Designing Resilient Systems Beyond Retries (Part 2): Bulkheading, Load Balancing, and Fallbacks
Designing Resilient Systems Beyond Retries (Part 1): Rate-Limiting
Context Deadlines and How to Set Them
Recipe for Building a Widget: How We Helped to “Peak-Shift” Demand by Helping Passengers Understand Travel Trends
Structured Logging: The Best Friend You’ll Want When Things Go Wrong
How We Simplified Our Data Ingestion & Transformation Process
Understanding Supply & Demand in Ride-hailing Through the Lens of Data
A Lean and Scalable Data Pipeline to Capture Large Scale Events and Support Experimentation Platform
Designing Resilient Systems: Circuit Breakers or Retries? (Part 2)
Querying Big Data in Real-time with Presto & Grab's TalariaDB
Designing Resilient Systems: Circuit Breakers or Retries? (Part 1)
Orchestrating Chaos Using Grab's Experimentation Platform
Reliable and Scalable Feature Toggles and A/B Testing SDK at Grab
Mockers - Overcoming Testing Challenges at Grab
Journey of a Tourist via Grab
How We Designed the Quotas Microservice to Prevent Resource Abuse
Grab Senior Data Scientist Liuqin Yang Wins Beale-Orchard-Hays Prize
Building Grab’s Experimentation Platform
Introducing Grab-Kit: Distributed Service Design at Grab
How Grab Experimented with Chat to Drive Down Booking Cancellations
Deep Dive into Database Timeouts in Rails
Dealing with the Meltdown Patch at Grab
GrabShare at the Intelligent Transportation Engineering Conference
Grabbing Growth: A Growth Hacking Story
The Data and Science Behind GrabShare Part I: Verifying Potential and Developing the Algorithm
The Art of Hiring Good Engineers
Migrating Existing Datastores
So You Need to Hire Good Engineers
Come and #hackallthethings at Grab
How We Scaled Our Cache and Got a Good Night's Sleep
Grab's Front End Study Guide
DNS Resolution in Go and Cgo
Driving Southeast Asia Forward with AWS
How to Go from a Quick Idea to an Essential Feature in Four Steps
Troubleshooting Unusual AWS ELB 5XX Error
Scaling Like a Boss with Presto
Deep Dive into iOS Automation at Grab - Continuous Delivery
Deep Dive into iOS Automation at Grab - Integration Testing
A Key Expired in Redis, You Won't Believe What Happened Next
How Grab Hires Engineers in Singapore
Battling with Tech Giants for the World's Best Talent
This Rocket Ain't Stopping - Achieving Zero Downtime for Rails to Golang API Migration
Grab Vietnam Careers Week
GrabPay Wins Best Fraud Prevention Innovation at the Florin Awards
Round-robin in Distributed Systems
Why Test the Design with Only 5 Users
Programmers Beware - UX is Not Just for Designers
Grab You Some Post-Mortem Reports
The Curious Case of the Phantom Instance
No articles match this topic.