As CRED grew, so did the invisible machinery behind it. Scaling from 100 to over 400 database instances pushed our monolithic Percona Monitoring Management (PMM) to its breaking point. This is the story of how we decoupled our observability stack, pivoting to VictoriaMetrics…
High-Risk, High-Scale: Guaranteeing Ad Budget Precision at 1 Million Events/Second
Flipkart serves thousands of sponsored ads across various pages and with search queries. Advertisers are charged based on the impression views or clicks generated by their campaign content. In the high-velocity domain of AdTech, the latency between an ad impression/click and a…
Inside Meesho’s Real-Time Analytics Platform: A Deep Dive into Driving Performance at Scale
At Meesho, we are constantly striving to understand our users better and enhance their experience. Every day, users can generate billions of events on our platform from browsing products to completing purchases and we aim to translate these signals into insights instantly. From…
Bringing Interfaces to Life with Rive: Real-Time Animations at slice
At slice, we’ve always believed that a good UI shouldn’t just react, it should feel alive. When we started building new experiences like Voice Pay, we quickly ran into a problem that traditional animation tools just couldn’t deliver the kind of real-time interactivity we were…
Keeping Java as the Core, Python to Lead Agentic Systems
We’ve been a Java shop for years, and now agentic AI demands us to rethink. How I got here? I’ve been in the Java world for more than 30 years. Java has been my default answer to “what should we build this in?” for most of my career. I’ve seen it grow from applets and early web…
Triksha — Securing AI with AI at Flipkart TL;DR GenAI apps introduce a new attack surface where prompts become the primary vector of exploitation, making generic application security ineffective for real-world, fine-tuned LLMs like e-commerce support bots. Triksha, Flipkart’s…
Context Ambitious goal and the scope of the problem We were chasing a critical, futuristic business requirement for our big data pipelines achieving “1-day planning,” defined as a total execution window of 24 hours or less. The entire process is extensive, processing a 7TB data…
Beyond the Context Window: A New Approach to Summarizing Big Data
Image Source: Google Nano BananaTL;DR LLMs struggle to summarize the thousands of reviews for a single product because of context window limits and conflicting opinions. The proposed solution, XL-OPSUMM, is a scalable framework that summarizes reviews incrementally (in chunks)…
Building slice’s Chatbot: An Engineering Journey (Series 2)
Tags:Intent Detection, Decision Tree, User Experience, Support Automation Introduction Modern customer support faces a fundamental paradox: despite having comprehensive user data, transaction histories, and real-time system states, most chatbot interfaces still present users…
Powering Real-Time Decisions with PhonePe’s Garfield Platform and Apache Druid
At PhonePe’s scale, handling millions of transactions and user interactions every hour requires an analytics platform that’s not just fast, but instant. When every millisecond counts — both for user experience and incident detection — batch processing simply doesn’t cut it.…
Centralized Dynamic Navigation at Scale: Zero Crashes, Module Independence, Faster Delivery
How we achieved a crash less centralised navigation system using deep-links and runtime destination loading in our android app. Pre-requisite: This is an advanced concept, Basic Android Navigation understanding is required: NavController: Manages navigation between…
Building Meesho's ML Platform: Scaling Model Inference & Real-Time Embedding Search (Part 3)
By mid-2023, we had transformed our ML stack—building a real-time feature store, optimizing model retrieval, and fine-tuning ranking. But two critical gaps remained: 🔹 Scaling model inference without hitting infrastructure roadblocks 🔹 Moving embedding search from batch to…
Elastic Spark: Autoscaling Structured Streaming at Sharechat
At ShareChat — a leading social media platform with over 200M monthly active users (MAU) — our data platform ingests and processes more than 1 million events per second. However, this event load isn't constant.
Building Meesho’s ML Platform: Lessons from the First-Gen System (Part 2)
By late 2022, we had built something we were truly proud of—a real-time ML serving system with a DAG-based executor, a feature store, and an interaction store powering key ranking and personalization models. It was a major milestone, the culmination of months of effort from data…
Realtime Content Deduplication At Scale: How We Cut Cost By 86%
Every time a user views a post, an event is fired to inform us. We feed this real‑time stream into our recommendation system, e.g. analytics pipelines, feature computations, and view counters.
The Creative Intelligence Unit (CIU): Meesho’s Playbook for Scalable, AI-First Creativity
At Meesho, creativity isn’t about out of the box ideas, it’s about building messages grounded in our user’s truth. We speak in the language of our users, shaped by insight, and designed to deliver impact. For us, creativity is a powerful loop. It begins with a user's insight,…
Evolving Search Ranking: From XGBoost to Multi-Task Deep Learning with Knowledge Distillation
Search at Meesho For an e-commerce platform like Meesho, optimizing Search is pivotal to enhancing the user experience, improving product visibility and faster discovery, and eventually increasing conversion rates. Users search in multiple Indian languages looking for quality…
Pincode Migration Journey with Compose Multiplatform
In today’s diverse technological landscape, the ability to create applications that seamlessly operate across multiple platforms is a significant advantage. At Pincode, our seller app, was initially created using Android SDKs with Compose UI in Kotlin, and catered exclusively to…
Revealing the Hidden Links: How We Break Down Complex Fraud Networks
The Rising Tide of E-commerce Fraud We are deeply aligned with our aim to democratize our platform—with new entrepreneurs joining and logistics partners working across pin codes in India. This inclusivity is our biggest strength—and also the reason why we need to be extra…
Scaling Deployments at Meesho: How We Built a Custom Rollout Service to Replace Our Existing System
Introduction: The Need for Bulletproof Deployments at Scale As our platform scaled and user expectations rose, so did the need for rock-solid infrastructure. One critical challenge surfaced: ensuring zero-downtime deployments while rolling out new features at speed. Despite…
Building Slice’s Chatbot: An Engineering Journey (Series 1)
Building slice’s Chatbot: An Engineering Journey (Series 1) Tags: MQTT, EMQX, Kafka, Real-Time System, Event-Driven Systems Introduction As our chatbot requirements grew, we transitioned from a managed chatbot solution to an in-house system. This blog highlights the challenges…
Personalization from Day One: Solving the Cold Start Problem at Meesho
New users rarely give second chances. If their first experience on the app, especially the homepage product feed, feels irrelevant then they leave quickly. At Meesho, we noticed that our static homepage product feed for new users was causing high bounce rates and a poor first…
Scaling Secret Detection at Meesho: Building a Pre-Commit Workflow with TruffleHog
The Challenge: Scaling Secure Development As engineering teams and codebases grow, managing sensitive information like API keys, tokens, and config files becomes increasingly challenging. These secrets can occasionally slip through, and resolving such issues like rotating…
Problem Statement PhonePe, as a payments platform, has an essential feature that allows users to check all past transactions. When payments are made to businesses/merchants, users have access to a historical view of all such transactions. For peer to peer transactions, however,…
What is Feed Re-ranking? Online shopping platforms face intense competition, where the ability to deliver the most relevant products to customers is crucial. A well-organized and relevant product feed, which inherently is a list of products, forms the backbone of any E-commerce…
Building Meesho’s ML Platform: From Chaos to Cutting-Edge (Part 1)
The Genesis: How a Friday Night Roast Sparked Meesho’s ML Platform It all started in early 2022, over a casual Friday evening catch-up. Like many great origin stories, this one began with friendly banter between a group of backend engineers and data scientists. As the…
Scaling Seamless Communication: How Meesho Powers 10B+ Messages Daily with Multi-Tenancy
Revolutionizing Communication at Scale At Meesho, seamless communication is at the heart of our user experience. Whether it’s an OTP for login, an order update, or a personalized shopping recommendation, our platform ensures timely and reliable messaging across multiple…
Scaling Seamless Communication: How Meesho Powers 10B+ Messages Daily with Multi-Tenancy
Revolutionizing Communication at Scale At Meesho, seamless communication is at the heart of our user experience. Whether it’s an OTP for login, an order update, or a personalized shopping recommendation, our platform ensures timely and reliable messaging across multiple…
The Journey to Faster Android Builds: How We Slashed Our Build Times at slice
As an Android developer at slice, watching our app evolve from a simple financial product to a full fledged bank has been exciting. But this growth came with a challenge familiar to many developers: increasingly painful build times. What started as a minor inconvenience…
In the early days of PhonePe, we started with just one application. Our pace was fast, almost like a race against time, and our initial engineers managed to keep things tidy. Colors, typography, and size constants were well-organized, ensuring a cohesive look and feel. Our small…
Android Network Observability: Shedding Light on Performance Blind Spots
Introduction In Android development, network performance issues can be notoriously hard to debug. While tools like Firebase Performance Monitoring provide high-level insights like error rates and latency, they fall short in providing granular visibility into each stage of the…
Seamlessly Migrating from MongoDB to PostgreSQL: Leveraging Data Platform for Scalable Success
Freaking hard Migration from MongoDB to PostgreSQL: Leveraging Data Platform for Scalable Success Introduction In today’s fast-paced digital landscape, selecting the right database technology is a crucial decision for any organisation’s success. Initially, MongoDB’s flexible…
DRS: Making high-impact & informed decisions with Dream11’s In-house Experimentation Platform!
With millions of Indian sports fans across our Dream Sports ecosystem, it takes a highly data-driven team to understand what is required for great fan engagement and ultimately making sports better through our various brands. At Dream11 too, we have been on a constant mission to…
Vulnerability Management & Risk Scoring for Business Services
In the rapidly evolving FinTech sector, PhonePe Private Limited, an India-born leader in financial services, stands out for its innovative range of products and services. Offering everything from UPI transactions and bill payments to stock market investments, personal finance…
Towards delivering world-class recommendations at the lowest possible cost
Sharechat is India’s largest homegrown social media company with native and regional user generated image and video content. Our two apps, ShareChat and Moj, are used by over 325 million users in 15 regional languages across India.
All an SRE needs to know : Automation ERA in Distributed Datastores
– Elasticsearch Upgrade for a cluster takes more than 10hrs by 2SREs manually and entails considerable back and forth. – SRE has automated the entire flow for rolling upgrades without downtime along with strict guardrails, killswitch, regular slack bots. – Automation has…
Virtual Machine Provisioning and Management in PhonePe
This blog post provides a high-level overview of PPeC (PhonePe Cloud) Agent’s architecture, its interaction with other key components like the PPeC API and PPeC Proxy, and how it optimizes tasks such as VM creation, resource allocation, and dynamic disk management. Evolution and…
Edge Framework: Powering On-Device ML for Smarter, Safer Experiences
In a world where every delivery is 2 minutes away, and every service is only a tap away, the need for devices to be smarter, faster, and more secure constantly is high. At the same time, users are more aware of their data privacy and expect seamless experiences without the lag…
This is written by Sarthak – Staff engineer, BankingThis write up summarizes some Opinions™️ that I have around good alerts, particularly in the context of infrastructure and platforms. The advice in here doesn’t represent universal truths, merely my experience over being…
Intro PhonePe’s apps have over 560m registered users, and are built by 250+ mobile engineers. The company ships 7 different apps (10, if you count Android and iOS separately). Operating at this scale, it’s been necessary to invest in and build a solid app platform. This blog…
How Urban Company leverages Causal Inference to power Data-Driven decisions
By — Kshitiz Swaroop ( Data Scientist, Data) In a world overflowing with data, finding correlations is easy. The real challenge is understanding causality. This article presents 2 case studies where the methodology was used to improve the organization wide estimates on the…
Discover what goes into a good GraphQL API for MongoDB, what it takes to build and maintain one, and how Hasura’s new framework allows developers to kickstart new projects and scale existing projects rapidly.
Demystifying TStore: The Backbone of Billions of Transactions at PhonePe – Chapter 2
Chapter 2 : A Deeper Dive In chapter 1, we explored TStore’s functionalities like high-velocity ingestion, efficient reads, and flexible querying, its core concepts like Entities, Units, and Views and finally components like the Client Bundle, Schema Registry, Entity Ingestor…
Demystifying TStore: The Backbone of Billions of Transactions at PhonePe
Chapter 1 : Unveiling TStore Introduction You are standing in the queue at a billing counter of a grocery store, and when your turn comes, you bill your items and make a scanned payment using the PhonePe app. Typically, you witness the green screen instantly and the latest…
LiquidUI – PhonePe’s Server-Driven UI Framework – Part 2
In our previous blog, we explored the numerous benefits of using Server-Driven UI (SDUI) over traditional native approaches. We also briefly discussed the three core components of LiquidUI – Server Driven UI Framework at PhonePe.
As PhonePe continues to grow rapidly, the scale and reliability of our infrastructure becomes more important. By developing our own infrastructure, we’re able to adapt quickly while keeping our services reliable. In order to meet our deployment needs, we use specialized in-house…