---
title: "Latest reads"
description: "The engineering internet, summarised so you can actually read it."
---

# Latest reads
> The engineering internet, summarised so you can actually read it.

## Articles

### [Streamlining Grab's Segmentation Platform with faster creation and lower latency](https://yomu.fyi/post/streamlining-grab-s-segmentation-platform-with-faster-creation-and-low.md)
- Company: [Grab](https://yomu.fyi/company/grab.md)
- Author: Jake Ng
- Published: Aug 15, 2023
- Type: Problem & solution

Grab's Segmentation Platform previously stored user-to-segment mappings across individual rows in ScyllaDB, causing write bottlenecks during segment creation and read latencies too high for downstream consumers. To resolve these performance limitations, the team transitioned to storing segments as Roaring Bitmaps saved as single blobs in object storage. This compression strategy splits 32-bit integer user IDs into chunks across array, bitmap, and run containers based on data density, reducing a one-million-member segment to under one megabyte. An accompanying client SDK manages segment retrieval, decoding, update notifications, and least-recently-used in-memory caching. Consequently, consumers such as Grab's communications platform achieved peak throughput of 15,000 queries per second with sub-millisecond p99 read latencies.


### [Coding the stars - an interactive constellation with Three.js and React Three Fiber](https://yomu.fyi/post/coding-the-stars-an-interactive-constellation-with-three-js-and-react.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: fsansalvadore
- Published: Aug 4, 2023
- Type: Tutorial

The article provides a tutorial for recreating Supabase Launch Week 8’s animated constellation in Next.js with Three.js and React Three Fiber (R3F). Each signup or user-generated ticket becomes a particle shaped into an “8,” with signups loaded from the \`lw8\_tickets\` table and later updates delivered through Supabase Realtime. The scene uses five-sided circle geometry, a white standard material with optional AdditiveBlending, an R3F Canvas, and a camera positioned at \`\[0, 0, 500\]\`. Particle movement is calculated with \`useFrame\`, elapsed time, sine and cosine functions, alternating loop direction, randomized speed and delay, and configurable offsets; a \`dat.gui\` debug mode exposes parameters such as particle count and shape dimensions.


### [pgvector: Fewer dimensions are better](https://yomu.fyi/post/pgvector-fewer-dimensions-are-better.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: gregnr,oli\_rice,egor\_romanov
- Published: Aug 3, 2023
- Type: Benchmark

Embeddings support semantic search, retrieval augmented generation, and clustering in pgvector, but high-dimensional vectors increase storage, memory, and distance-computation costs as datasets grow. The post explains how IVF indexes reduce the records examined and compares embedding models using Massive Text Embedding Benchmark results, noting little correlation between dimension count and performance. In a 224,482-vector Wikipedia dataset, all-MiniLM-L6-v2 with 384 dimensions delivered 78% higher queries per second than OpenAI text-embedding-ada-002 with 1536 dimensions at accuracy@10 of 0.99, while using 4GB rather than 7.5GB of RAM. The 384-dimensional gte-small model produced more than a 200% queries-per-second improvement over all-MiniLM-L6-v2 at the same accuracy, though model choice still depends on task performance, sequence length, language support, and operational costs.


### [Supabase Beta July 2023](https://yomu.fyi/post/supabase-beta-july-2023.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: ant\_wilson
- Published: Aug 2, 2023
- Type: Announcement

Supabase’s July 2023 update lists product releases, community content, and preparations for Launch Week 8, which was scheduled to begin the following week. The shipped changes include a new Android Kotlin client library, Login with Figma as an OAuth provider, Resend support as a custom SMTP provider, pgsodium 3.1.8, and PostgREST 11.1. A performance benchmark examines pgvector 0.4.0’s strengths and limitations, while noting that work on HNSW had started and was not far away. The update also features a Snowflake natural-language query application built with Python, Streamlit, Open AI, and Supabase Vector, plus tutorials, courses, community projects, and hiring announcements.


### [Unsupervised graph anomaly detection - Catching new fraudulent behaviours](https://yomu.fyi/post/unsupervised-graph-anomaly-detection-catching-new-fraudulent-behaviour.md)
- Company: [Grab](https://yomu.fyi/company/grab.md)
- Author: Rizal Fathony
- Published: Aug 2, 2023
- Type: Problem & solution

Fraud detection systems relying on historical training data often struggle to detect new fraudulent modus operandi because emerging patterns lack labeled supervision. Grab developed GraphBEAN, an unsupervised autoencoder model designed to detect anomalous patterns across bipartite interaction graphs representing consumers and merchants. The architecture uses graph convolution layers to encode node and edge attributes, then reconstructs features and edge existence through separate feature and structure decoders. High reconstruction errors flag rare, anomalous behaviors at both the node and edge levels. An automated pipeline applies heuristic fraud tags to these scores and routes them to human analysts and automated mitigation systems.


### [What I learned taking Ramp Bill Pay from 0 to N](https://yomu.fyi/post/what-i-learned-taking-ramp-bill-pay-from-0-to-n.md)
- Company: [Ramp](https://yomu.fyi/company/ramp.md)
- Author: Pavel Asparouhov
- Published: Aug 1, 2023
- Type: Explainer

An engineer recounts lessons from taking Ramp’s Bill Pay from its first code in April 2021 to a publicly launched product that eventually became a product with multibillion annual total payment volume. For 0-to-1 work, the account emphasizes protecting core abstractions, including canonicalized payment statuses that reduce varied payment states to five, using mocked functions to coordinate parallel engineering, and reaching beta customers quickly. Customer feedback, session replay, and persistent iteration are presented as necessary because the first customer will usually expose substantial flaws. For 1-to-N growth, manual processes and duct tape must be prioritized for replacement according to user impact, error risk, and workload. Growing stakes also require testing, monitoring, contingency plans, gradual rollouts, QA, and documentation that helps engineers self-service onboarding.


### [React Native file upload with Supabase Storage](https://yomu.fyi/post/react-native-file-upload-with-supabase-storage.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: simon\_grimm
- Published: Aug 1, 2023
- Type: Tutorial

This tutorial shows how to build a React Native image-sharing or file-sharing app with Expo SDK49, Supabase Authentication, Expo Router v2, and Supabase Storage. It configures a Supabase project, turns off email confirmation for the tutorial, creates a private files bucket, and adds policies that constrain access to an authenticated user’s own folder. The React Native client stores sessions in AsyncStorage, uses the Image Picker and Expo File System, and initializes Supabase with environment-based URL and public key values. The resulting app supports account creation, sign-in, image selection, listing, download-based display, and deletion; resumable uploads are omitted because the attempted implementation produced files with zero bytes.


### [Sidekick’s Improved Streaming Experience](https://yomu.fyi/post/sidekick-s-improved-streaming-experience.md)
- Company: [Shopify](https://yomu.fyi/company/shopify.md)
- Author: 2023-10-18
- Published: Aug 1, 2023
- Type: Explainer

Sidekick addresses two streaming problems in LLM chatbots: Markdown fragments can render awkwardly before syntax is complete, and tool-driven responses can delay visible output across multiple LLM roundtrips. Its solution combines a buffering Markdown parser with an event emitter that multiplexes the main response and asynchronously resolved content into one stream. A Node.js Transform stream processes Unicode characters through a finite-state machine, buffering ambiguous sequences such as emphasis, list markers, and links until they resolve or are flushed as raw text. The parser also recognizes card links, triggering asynchronous card-resolution tasks while Server-Sent Events carry named streams to the UI, which demultiplexes them into the main response and placeholders. The design uses Markdown as a human-readable transport between the LLM, backend, and frontend, allowing content to appear as it is produced.


### [Enterprise](https://yomu.fyi/post/enterprise.md)
- Company: [Okta](https://yomu.fyi/company/okta.md)
- Author: Semona Igama
- Published: Jul 28, 2023
- Type: Tutorial

This workshop shows SaaS developers how to add SCIM provisioning to an application so enterprise customers can manage users across systems, including people who have not yet logged in. It builds SCIM server functionality into a React and Express Todo application, connects it to an Okta client, and tests lifecycle operations such as importing, updating, deprovisioning, and reprovisioning users. The implementation follows SCIM 2.0 with JSON-based REST endpoints for creating, retrieving, filtering, updating, and deleting users, plus bearer-token authentication, SCIM content-type headers, logging, and soft deletion. It also explains production considerations, including unique identifiers and mapping each request’s API token to the correct organization for multitenancy. The completed application supports OIDC authentication alongside SCIM provisioning.


### [Enterprise Maturity Workshop: Terraform](https://yomu.fyi/post/enterprise-maturity-workshop-terraform.md)
- Company: [Okta](https://yomu.fyi/company/okta.md)
- Author: E. Dunham
- Published: Jul 28, 2023
- Type: Tutorial

This workshop teaches how to use Okta’s Terraform provider to manage users and groups in an Okta Organization while practicing beginner and intermediate Terraform skills. It explains desired-state configuration, where terraform plan predicts changes and terraform apply makes them, and describes how providers, resources, variables, data sources, and the terraform.tfstate file fit into a project. The exercises cover provider setup with a client ID and private key, secret handling, creating and assigning users and groups, importing resources, refactoring files, formatting code, and controlling API usage. It also uses errors and configuration drift to show how Terraform and Okta can disagree, then compares importing manual changes with applying Terraform’s declared state. The workshop concludes with cleanup through terraform destroy and points to further provider and Terraform documentation.


### [How to Get Going with the On](https://yomu.fyi/post/how-to-get-going-with-the-on.md)
- Company: [Okta](https://yomu.fyi/company/okta.md)
- Author: Alisa Duncan
- Published: Jul 27, 2023
- Type: Tutorial

This workshop setup guide prepares a minimal Todo SaaS application for a series of enterprise-readiness workshops, while warning that the base app is not production quality and has intentional security shortcomings. It uses a React frontend, an Express API written in TypeScript, SQLite persistence, Prisma for ORM and database visualization, and an Nx workspace for shared project commands. Developers install Node.js v18 or higher, obtain the repository, run npm ci, seed two users with npm run init-db, and start the application with npm start. The frontend runs on localhost:3000, the API on localhost:3333, and Prisma Studio on localhost:5555 for inspecting users and todos. The guide concludes by directing developers to workshops covering OpenID Connect, SCIM, Terraform, Okta Workflows, and cross-application sign-out.


### [Shopify’s platform is the Web platform](https://yomu.fyi/post/shopify-s-platform-is-the-web-platform.md)
- Company: [Shopify](https://yomu.fyi/company/shopify.md)
- Author: 2023-10-18
- Published: Jul 26, 2023
- Type: Announcement

Shopify now recommends Remix for building Admin apps, which merchants use to configure apps, view data, and integrate services within Shopify Admin. Because these apps run in cross-origin iframes, browser restrictions can block cookies, storage, URL updates, and direct access to the Admin interface; App Bridge uses postMessage() to restore capabilities and support deeper integration. The new App Bridge client replaces the previous package, requires only a script tag, and makes standard web APIs such as history.pushState() work with Admin apps. The Remix template uses file-based routes, server-side loaders and actions, implicit API handling, and Shopify-specific integrations supplied by @shopify/shopify-app-remix, with Prisma provided for storage and sessions. Shopify integrated the template into Shopify CLI and describes it as the beginning of broader platform and design-system reworks.


### [Supabase Launch Week 8 Hackathon](https://yomu.fyi/post/supabase-launch-week-8-hackathon.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: ant\_wilson
- Published: Jul 25, 2023
- Type: Announcement

Supabase Launch Week 8 will run an asynchronous hackathon for new Open Source projects that use Supabase in some capacity. The event is listed as starting at 09:00 am PT on Friday 4 August 2023, with the main schedule ending on Sunday 13 August at 11:59 pm PT; the submission instructions separately give a deadline of 11:59 pm PT on Sunday 14 August 2023. Individuals or teams of up to five can build an app, tool, library, or other project in any language or framework, with all design and code created during the event. Judges will award winners and runners-up across five categories, including overall quality, AI use, technical impressiveness, visual appeal, and fun, while submissions must include source code and may earn extra points for a demo video.


### [7 Steps to a Complete Social Media Audit](https://yomu.fyi/post/7-steps-to-a-complete-social-media-audit.md)
- Company: [Brandwatch](https://yomu.fyi/company/brandwatch.md)
- Author: Gemma Joyce
- Published: Jul 21, 2023
- Type: Tutorial

Social media audits are presented as a way for marketers to connect channel activity with business goals and improve return on investment, especially when a small share of content or actions drives most results. The guide organizes the work into seven steps: reviewing analytics and presence, setting goals, analyzing performance, understanding the audience, studying competitors, and creating an action plan. It recommends consolidating channel data with a social media management platform or native analytics, using social listening to examine conversations, interests, and sentiment, and evaluating metrics such as engagement, reach, posting timing, website behavior, and conversions. The process ends with documented findings, team discussion, and prioritized changes aligned with company objectives, with an annual audit recommended.


### [Multiple Logs for Resiliency](https://yomu.fyi/post/multiple-logs-for-resiliency.md)
- Company: [Fly.io](https://yomu.fyi/company/fly-io.md)
- Published: Jul 21, 2023
- Type: Problem & solution

An unexpected network outage can make an application misbehave while also removing crucial log data from the normal Fly.io pipeline. The post traces that pipeline from application STDOUT through Fly.io, NATS, the Log Shipper, Vector, and a third-party provider, then presents two ways to add redundancy. One strategy broadcasts framework logs to a separate file on each machine, preferably on a persistent volume with rotation; a Rails example uses ActiveSupport::Logger.broadcast. The other runs a separate app that subscribes directly to NATS, parses messages, and writes selected fields to disk, with a Fly authentication token and provisions for reconnection and error handling. File copies are more failsafe but raw and harder to fetch, while a custom shipper can produce more usable data but is less bulletproof.


### [Inference characteristics of Llama](https://yomu.fyi/post/inference-characteristics-of-llama.md)
- Company: [Cursor](https://yomu.fyi/company/cursor.md)
- Author: Aman Sanger
- Published: Jul 20, 2023
- Type: Benchmark

Llama-2-70B is evaluated against GPT-3.5-turbo on serving cost and latency, with the comparison deliberately excluding GPT-4 and noting fine-tuning as a separate reason to choose Llama. The evaluation serves Llama on two 80-GB A100 GPUs in 16-bit precision, then uses transformer FLOP, memory-bandwidth, and KV-cache analysis to explain the difference between prompt processing and token generation. The measurements indicate that Llama-2 has worse completion-token pricing than GPT-3.5 on two A100s, while its prompt-token pricing is more than three times cheaper; eight A100s might improve pricing but introduce unacceptable latency. The resulting guidance favors GPT-3.5 for most workloads and open-source Llama models for prompt-heavy tasks such as classification or reranking, with batch processing and workloads lacking prompt tokens identified as additional cases where Llama may make sense.


### [Getting started with Flutter authentication](https://yomu.fyi/post/getting-started-with-flutter-authentication.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: tyler\_shukert
- Published: Jul 18, 2023
- Type: Tutorial

This tutorial shows how to add Google authentication to a cross-platform Flutter application using the Supabase SDK for Flutter. It explains OpenID Connect as a protocol built on OAuth 2.0, then configures Google Cloud client IDs for iOS, Android, and web, with Supabase using the web client ID to verify the supplied identity token. The sample initializes Supabase, starts Google sign-in through google\_sign\_in, passes the resulting access and ID tokens to signInWithIdToken, and listens for signed-in events to navigate from LoginScreen to ProfileScreen. The profile reads avatar\_url and full\_name from Supabase userMetadata, while a sign-out action returns the user to the login screen, producing a basic personalized app.


### [Contributing support for a Wasm instruction to Winch](https://yomu.fyi/post/contributing-support-for-a-wasm-instruction-to-winch.md)
- Company: [Shopify](https://yomu.fyi/company/shopify.md)
- Author: 2023-10-18
- Published: Jul 18, 2023
- Type: Tutorial

Adding a WebAssembly instruction to Winch, Wasmtime’s baseline non-optimizing compiler, requires coordinated changes across code generation, register management, ISA-specific emission, and tests. The contribution described adds visitor support, delegates through CodeGenContext and MacroAssembler, emits x86\_64 instructions through Cranelift-backed assembler methods, and registers the instruction for differential fuzzing. CodeGenContext keeps the value stack, register allocator, and function frame consistent, while spilling can move values to linear-memory offsets; MacroAssembler handles ISA differences, register constraints, and CPU feature fallbacks such as lzcnt versus bsr-based clz. Filetests inspect generated machine code without executing it, while differential fuzzing compares Winch results with another WebAssembly engine across randomly generated modules and inputs. The workflow also includes running fuzzing, incorporating draft-PR feedback, rebasing, and submitting the public Wasmtime pull request.


### [Moving Fast by Moving Slow: How We Built Payments at Ramp](https://yomu.fyi/post/moving-fast-by-moving-slow-how-we-built-payments-at-ramp.md)
- Company: [Ramp](https://yomu.fyi/company/ramp.md)
- Author: Shreyan Jain
- Published: Jul 17, 2023
- Type: Problem & solution

Ramp describes why it delayed the launch of BillPay by about two months to replace its existing, ad hoc bank-transfer infrastructure with a new payments service. The earlier system encoded product-specific concepts such as employee bank accounts and reimbursement transfers, forcing each new use case to add bespoke data models, jobs, and scheduling logic. This friction became especially costly as the team anticipated additional payment products and vendor payments scheduled for future dates. The redesign used generic primitives—ACHDetails and TransferCanonical—to separate business logic from money-movement logic; it also treats failed transfers and retries as concerns requiring an auditable identity, history, and consistent provider mapping. The post argues that stress-testing assumptions, modeling relevant degrees of freedom, limiting conditional logic, and building platforms after recurring patterns can make future products faster to launch.


### [pgvector 0.4.0 performance](https://yomu.fyi/post/pgvector-0-4-0-performance.md)
- Company: [Supabase](https://yomu.fyi/company/supabase.md)
- Author: egor\_romanov,pavel
- Published: Jul 13, 2023
- Type: Benchmark

The post benchmarks pgvector 0.4.0 for approximate nearest-neighbor search and assesses whether PostgreSQL can support production-scale vector workloads. Using the ANN Benchmarks methodology, pre-warmed RAM, the dbpedia-entities-openai-1M dataset of 1 million 1,536-dimensional OpenAI embeddings, and Supabase ARM instances, it varies probes, lists, concurrency, and compute size. On an 8-core, 32 GB instance, pgvector reached 0.91 accuracy@10 at 380 QPS with 10 probes and 0.98 accuracy at 140 QPS with 40 probes; a 64-core, 256 GB server reached about 1,800 QPS at 0.91 accuracy. The results indicate predictable scaling, while higher lists can improve query speed at the cost of index-build time, and the roadmap identifies HNSW, product quantization, and parallel index builds as planned improvements.


[Newer posts](https://yomu.fyi/posts/page/192.md) · [Older posts](https://yomu.fyi/posts/page/194.md)