Sahil Kapoor's Playbook

On-the-ground startup lessons, product frameworks, and engineering insights from the trenches. For builders who move fast and fix scale as it breaks. 🌉 bridged from ⁂ https://sahilkapoor.com/, follow @ap.brid.gy to interact

6 followers1 following169 stories

Longform Stories

Webpack

Webpack is a JavaScript module bundler that takes a project's source files (JS, CSS, images, fonts) and produces optimised static assets for the browser. Webpack defined the modern front-end build pip…

May 20·1 min read·81 words

Vue.js

Vue.js is a progressive JavaScript framework for building user interfaces, created by Evan You in 2014. Vue is designed to be incrementally adoptable: it can drop into a single page or anchor an entir…

May 19·1 min read·78 words

What is NVIDIA DGX Spark (and why I bought one)

NVIDIA DGX Spark is a desktop AI supercomputer with 128GB unified memory. Here's what it is, how it compares to a Mac or consumer GPU, and why I bought one to dive deeper into AI.

May 18·7 min read·1340 words

Vite

Vite is a front-end build tool created by Evan You (the creator of Vue), designed to give developers near-instant dev server starts and hot module replacement. It uses native ES modules and esbuild du…

May 18·1 min read·77 words

Helm

Helm is to Kubernetes what apt is to Ubuntu or npm is to Node.js. Instead of maintaining dozens of raw Kubernetes YAML files for every application, you package them into a Helm Chart with templated va…

May 17·3 min read·420 words

ArgoCD

ArgoCD is a Kubernetes controller that continuously watches a Git repository and ensures your cluster's actual state matches what's declared in that repo. If someone manually modifies a Kubernetes res…

May 17·3 min read·448 words

Pulumi

Pulumi is an Infrastructure as Code platform that challenges Terraform's HCL-based approach by letting you write infrastructure definitions in languages you already know. Instead of learning a new DSL…

May 17·3 min read·432 words

HashiCorp Vault

HashiCorp Vault is the de facto secrets management solution for production infrastructure. Instead of storing database passwords in environment variables, AWS keys in CI/CD pipelines, and API tokens s…

May 17·3 min read·439 words

Caddy

A modern web server and reverse proxy written in Go that provisions and renews TLS certificates automatically — the simplest path to HTTPS for self-hosted services.

May 17·2 min read·358 words

Traefik

Traefik (pronounced "traffic") is a reverse proxy built for dynamic environments. Where Nginx requires you to restart the process after every config change, Traefik watches your Docker socket or Kuber…

May 17·3 min read·434 words

Cloudflare Tunnel

Cloudflare Tunnel (formerly Argo Tunnel) creates an outbound-only encrypted connection from your server to Cloudflare's edge. Your origin server makes the connection out; Cloudflare routes incoming re…

May 17·3 min read·462 words

WireGuard

WireGuard is a VPN protocol created by Jason Donenfeld and merged into the Linux kernel in 2020. It replaces OpenVPN and IPsec for most new deployments with a dramatically simpler design: cryptokey ro…

May 17·3 min read·435 words

Tailscale

Tailscale takes the cryptographic strength of Wireguard and wraps it in a coordination layer that eliminates all the manual configuration that makes WireGuard hard to operate at scale. Install the cli…

May 17·3 min read·475 words

Inference Endpoint

An inference endpoint is the serving layer for a trained model. After training (or downloading) an LLM, you need infrastructure to accept requests, run the forward pass, and return outputs at scale. T…

May 17·3 min read·438 words

Tokenization

Tokenization is the first step in any LLM pipeline: converting raw text into a sequence of integer IDs that the model actually processes. Understanding tokenization helps you reason about context wind…

May 17·3 min read·469 words

RLHF (Reinforcement Learning from Human Feedback)

RLHF is the training recipe that turned raw language models (good at predicting text) into aligned assistants (good at following instructions helpfully and safely). It was popularized by InstructGPT (…

May 17·3 min read·459 words

LoRA (Low-Rank Adaptation)

LoRA (Low-Rank Adaptation) is a fine-tuning method introduced by Hu et al. at Microsoft in 2021. Instead of updating all billions of parameters in a large model, LoRA freezes the original weights and …

May 17·3 min read·512 words

System Prompt

In the chat completions API format used by OpenAI, Anthropic, Google, and others, messages come in three roles: system, user, and assistant. The system message is the system prompt, it's processed fir…

May 17·3 min read·483 words

Prompt Engineering

Prompt engineering is the discipline of communicating effectively with large language models. Because LLMs are trained to predict plausible continuations of text, how you frame a request has an enormo…

May 17·3 min read·474 words

LangChain

LangChain is an open-source framework that provides building blocks for LLM applications. Rather than calling OpenAI's API directly and wiring everything by hand, LangChain gives you composable abstra…

May 17·3 min read·470 words

OpenRouter

A unified API gateway for large language models that lets you call 100+ LLMs from different providers through a single OpenAI-compatible endpoint with automatic fallback and cost routing.

May 17·2 min read·339 words

OpenHands

OpenHands (formerly OpenDevin) is an open-source platform for AI software engineering agents. Unlike Cursor or Windsurf which are IDEs with AI assistance, OpenHands is a platform where AI agents opera…

May 17·3 min read·427 words

Windsurf

Windsurf (by Codeium) is an AI IDE that competes directly with Cursor. Its key differentiator is Cascade, an agentic system designed around "flows", coherent multi-step coding sessions where the agent…

May 17·2 min read·369 words

GitHub Copilot

GitHub Copilot, launched in 2021 and built on OpenAI Codex (later GPT-4), was the first AI pair programmer to reach mainstream adoption. It integrates as an extension into VS Code, JetBrains, Neovim, …

May 17·3 min read·455 words

Cursor

Cursor is a fork of VS Code that reimagines the editor as an AI-native environment. While GitHub Copilot adds AI as a plugin to existing editors, Cursor was built from the ground up with AI as a first…

May 17·3 min read·485 words

vLLM

vLLM (Virtual LLM) is an open-source inference engine from UC Berkeley that dramatically increases the throughput of serving large language models on GPU hardware. It was introduced in 2023 with Paged…

May 17·3 min read·464 words

Ollama

Ollama makes running open-source LLMs as straightforward as running a Docker container. You pull a model, and it starts serving a local REST API that your code can call, no cloud, no API key, no per-t…

May 17·3 min read·454 words

MCP (Model Context Protocol)

Model Context Protocol (MCP) is an open standard introduced by Anthropic in late 2024 that defines a universal way for large language models to communicate with external systems, files, databases, API…

May 17·3 min read·530 words

TypeScript

TypeScript is a typed superset of JavaScript developed by Microsoft. It adds static types, type inference, generics, interfaces, enums, and structural typing on top of standard JavaScript, then compil…

May 17·1 min read·74 words

Tailwind CSS

Tailwind CSS is a utility-first CSS framework that gives developers a large set of low-level utility classes (flex, pt-4, text-lg, bg-slate-900) which compose directly in HTML / JSX, instead of writin…

May 16·1 min read·77 words

Svelte

Svelte is a JavaScript UI framework that shifts most of the work from runtime to build time. Svelte components are compiled into small, surgical, imperative JavaScript that updates the DOM directly, r…

May 15·1 min read·78 words

Server-Side Rendering (SSR)

Server-Side Rendering (SSR) is the practice of rendering a page's HTML on the server in response to a request, then sending the fully formed markup to the browser, optionally followed by JavaScript th…

May 14·1 min read·81 words

Remix

Remix is a React framework focused on web-fundamentals: HTML forms, progressive enhancement, nested routing, and tight integration with the web platform. Remix moved into the React Router project in 2…

May 13·1 min read·79 words

Push Notifications

Push notifications are short messages that a backend server delivers to a user's device through the OS-level push service, displayed even when the corresponding app is not running. Push is the standar…

May 12·1 min read·86 words

PWA

A Progressive Web App (PWA) is a web application that uses modern browser capabilities to behave more like a native app: installable to the home screen, offline-capable, sending push notifications, an…

May 12·1 min read·73 words

Jetpack Compose

Jetpack Compose is Google's declarative UI framework for Android, built around Kotlin and shipped stable in 2021. Like SwiftUI on Apple platforms, Compose lets developers describe the UI as a function…

May 12·1 min read·76 words

SwiftUI

SwiftUI is Apple's declarative UI framework for building user interfaces across iOS, iPadOS, macOS, watchOS, tvOS, and visionOS. Introduced in 2019, SwiftUI uses Swift's strong type system and result …

May 12·1 min read·76 words

Flutter

Flutter is Google's cross-platform UI framework for building applications from a single Dart codebase. Unlike React Native, Flutter does not use platform-native UI widgets; it renders its own widgets …

May 12·1 min read·75 words

React Native

React Native is a cross-platform framework from Meta that lets developers build iOS and Android applications using React. UI components map to native platform widgets at runtime, so the apps feel nati…

May 12·1 min read·82 words

Kotlin

Kotlin is a statically typed, modern programming language developed by JetBrains and adopted by Google in 2017 as the preferred language for Android development. Kotlin runs on the JVM (with full Java…

May 12·1 min read·76 words

Swift

Swift is Apple's modern programming language, introduced in 2014 as the successor to Objective-C for building iOS, iPadOS, macOS, watchOS, tvOS, and visionOS applications. Swift is statically typed, m…

May 12·1 min read·68 words

Android

Android is Google's open-source mobile operating system, powering the majority of the world's smartphones across hundreds of device manufacturers. Android is based on a modified Linux kernel and ships…

May 12·1 min read·74 words

iOS

iOS is Apple's mobile operating system, powering the iPhone, iPad (as iPadOS, a close derivative), and iPod touch. iOS is a closed source platform with first-party SDKs, a curated App Store as the dis…

May 12·1 min read·75 words

Availability Zone

An availability zone (AZ) is an isolated data centre within a cloud region. AZs in the same region share low-latency private network links but use physically separate buildings, power, cooling, and up…

May 12·1 min read·83 words

Region

A region in cloud computing is a geographic area in which a cloud provider operates a cluster of data centres. Regions are independent of each other: an outage in one does not affect another, services…

May 12·1 min read·81 words

EC2

Amazon EC2 (Elastic Compute Cloud) is AWS's primary virtual machine service. EC2 provisions virtual servers (instances) in minutes, billed by the second, with a wide selection of CPU, memory, GPU, and…

May 12·1 min read·78 words

Lambda

AWS Lambda is Amazon's serverless function-as-a-service (FaaS). It runs short-lived functions in response to events, scales transparently from zero to thousands of concurrent invocations, and bills on…

May 12·1 min read·77 words

S3

Amazon S3 (Simple Storage Service) is AWS's object storage service. It stores arbitrary blobs of data (objects) in named containers (buckets), accessed over HTTPS via the S3 API. S3 is the original AW…

May 12·1 min read·86 words

Redis Cluster

Redis Cluster is Redis's native sharding mode. A cluster of Redis nodes partitions the keyspace into 16,384 hash slots; each node owns a contiguous range of slots; clients learn which node owns which …

May 12·1 min read·83 words

Valkey

Valkey is an open-source, BSD-licensed fork of Redis created in 2024 after Redis Ltd relicensed the Redis source from BSD to a dual SSPL/RSALv2 model that the Linux Foundation, AWS, Google Cloud, Orac…

May 12·1 min read·82 words

In-memory Database

An in-memory database keeps its primary dataset in RAM rather than on disk, trading capacity and durability tradeoffs for orders-of-magnitude lower read and write latency. In-memory engines back almos…

May 12·1 min read·69 words

PostGIS

PostGIS is the geospatial extension for PostgreSQL. It adds geometry and geography data types, hundreds of spatial functions, spatial indexes (GiST, SP-GiST, BRIN), and conformance to the OGC Simple F…

May 12·1 min read·73 words

pgvector

pgvector is the popular PostgreSQL extension that adds a vector column type, distance operators, and approximate nearest-neighbour indexes for similarity search over embeddings. It turns any PostgreSQ…

May 12·1 min read·70 words

Replication

Database replication is the practice of keeping multiple copies of the same data on different nodes for availability, durability, and read scaling. Almost every production database runs with at least …

May 12·1 min read·69 words

Replica Set

A replica set is MongoDB's name for a group of database nodes that maintain the same data and provide automatic failover. One node is the primary (accepts writes); the rest are secondaries that asynch…

May 12·1 min read·81 words

Sharding

Sharding is the practice of horizontally partitioning a dataset across multiple database instances so each shard holds a subset of the data and serves a subset of the load. Sharding is the standard an…

May 12·1 min read·87 words

Aggregation Pipeline

The Aggregation Pipeline is MongoDB's framework for transforming and combining documents through a sequence of stages, conceptually similar to a SQL SELECT with GROUP BY, JOIN, and window functions. A…

May 12·1 min read·81 words

BSON

BSON (Binary JSON) is the binary-encoded serialization format MongoDB uses to store and transmit documents. It extends JSON with additional data types (date, ObjectId, decimal128, binary data, regex),…

May 12·1 min read·72 words

Document Database

A document database stores data as self-contained documents (typically JSON or BSON) grouped into collections, instead of as rows in tables with strict schemas. Each document carries its own structure…

May 12·1 min read·79 words

NoSQL

NoSQL is an umbrella term for databases that depart from the strict relational, SQL-based, table-and-row model. The category emerged in the late 2000s as web-scale applications needed horizontal scali…

May 12·1 min read·69 words

OLTP

OLTP (Online Transaction Processing) describes the class of database workloads characterised by many short-lived, latency-sensitive transactions: row-level reads and writes that back interactive appli…

May 12·1 min read·66 words

MVCC

Multi-Version Concurrency Control (MVCC) is a concurrency mechanism in which writers do not block readers and readers do not block writers. Instead of locking rows for the duration of a transaction, t…

May 12·1 min read·81 words

ACID

ACID is the set of guarantees a database transaction provides: Atomicity, Consistency, Isolation, Durability. ACID is the bedrock contract that makes relational databases trustworthy for financial, in…

May 12·1 min read·73 words

MySQL

MySQL is one of the most widely deployed open-source relational databases, especially common in web stacks and as the storage layer behind countless WordPress, Rails, and PHP applications. It is owned…

May 12·1 min read·74 words

SQL

SQL (Structured Query Language) is the declarative language for defining, querying, and manipulating relational databases. Despite decades of speculation about its replacement, SQL remains the dominan…

May 12·1 min read·65 words

Pub/Sub

Pub/Sub (Publish-Subscribe) is a messaging pattern in which producers publish messages to a named channel (topic) without knowing who consumes them, and consumers subscribe to topics they care about w…

May 12·1 min read·73 words

Service Mesh

A service mesh is an infrastructure layer that handles service-to-service communication for a microservices application: routing, retries, timeouts, mTLS encryption, authorization, traffic shifting, a…

May 12·1 min read·74 words

Helm

Helm is the package manager for Kubernetes. It packages a set of Kubernetes manifests into a chart, parameterised by values, that can be installed, upgraded, rolled back, and shared via repositories. …

May 12·1 min read·71 words

Container Runtime

A container runtime is the software component that actually creates and runs containers from an OCI image. It manages namespaces, cgroups, mount points, capabilities, and the container process lifecyc…

May 12·1 min read·74 words

Image Layer

An image layer is one immutable filesystem diff that, when stacked with the layers above and below it, forms a complete container image. Each Dockerfile instruction (FROM, COPY, RUN, ADD) typically pr…

May 12·1 min read·81 words

Podman

Podman is an open-source container engine compatible with the Docker CLI but architected without a long-running root daemon. Containers run as a regular user, in a rootless namespace, with a pod abstr…

May 12·1 min read·78 words

Docker Compose

Docker Compose is a tool for defining and running multi-container applications on a single host using a declarative YAML file. It is the standard way to spin up a local development environment with a …

May 12·1 min read·75 words

BuildKit

BuildKit is the modern build backend for Docker and the OCI ecosystem. It replaces the original Docker builder with a concurrent, cache-aware engine that builds OCI images from Dockerfiles much faster…

May 12·1 min read·76 words

OCI

The Open Container Initiative (OCI) is an open governance body, hosted by the Linux Foundation, that maintains the standard specifications for container images, runtimes, and distribution. OCI specs a…

May 12·1 min read·74 words

containerd

containerd is an industry-standard container runtime daemon that manages the complete container lifecycle on a host: image pull and storage, container execution, snapshotting, and networking. It is th…

May 12·1 min read·72 words

Serverless

Serverless is a cloud execution model in which the provider runs and scales the underlying compute, and the customer is billed only for actual usage (request count, execution time, memory). Despite th…

May 12·1 min read·72 words

Cloud Computing

Cloud computing is the model of delivering compute, storage, networking, and software as on-demand services over the internet, billed by usage rather than by capital purchase. Cloud providers operate …

May 12·2 min read·273 words

CAP Theorem

The CAP theorem states that a distributed data store cannot simultaneously provide all three of: Consistency (every read returns the latest write), Availability (every request receives a response), an…

May 12·1 min read·72 words

Eventual Consistency

Eventual consistency is the property that, given enough time without new updates, all replicas of a piece of data will converge to the same value. Reads may see stale or out-of-order data in the short…

May 12·1 min read·77 words

Consensus

Consensus is the problem of getting a set of distributed processes to agree on a single value despite failures, message delays, and out-of-order delivery. Consensus protocols are the foundation under …

May 12·1 min read·72 words

Saga

A Saga is a pattern for executing a business transaction that spans multiple services or databases by chaining local transactions together, with compensating actions to roll back partial progress on f…

May 12·1 min read·78 words

Event-Driven Architecture

Event-Driven Architecture (EDA) is a style in which services communicate by publishing and subscribing to events on a message bus, rather than calling each other synchronously. Producers emit events w…

May 12·1 min read·78 words

Modular Monolith

A modular monolith is a single-deployable application built with strict internal module boundaries: clear ownership per module, explicit interfaces between them, no shared mutable state across boundar…

May 12·1 min read·73 words

Monolith

A monolith is an application packaged and deployed as a single unit: one codebase, one build, one process, one database. Everything ships together. Most applications start as monoliths and many remain…

May 12·1 min read·85 words

Pipeline

A pipeline is the automated sequence of stages a code change passes through on its way from a developer's commit to production. Modern CI/CD pipelines codify build, test, security scan, package, and d…

May 12·1 min read·80 words

Feature Flag

A feature flag (also called a feature toggle) is a runtime switch that turns a code path on or off without deploying new code. Flags let teams decouple deploy from release, ship dark, run experiments,…

May 12·1 min read·88 words

Canary Deployment

Canary Deployment is a release strategy that routes a small fraction of traffic to a new version first, observes its behaviour against the old version, and gradually expands the new version's share if…

May 12·1 min read·88 words

Blue-Green Deployment

Blue-Green Deployment is a release strategy that maintains two identical production environments and switches all traffic from one to the other when a new version is ready. The currently live environm…

May 12·1 min read·80 words

ArgoCD

ArgoCD is an open-source declarative continuous-delivery tool for Kubernetes that runs inside the cluster and applies manifests from a Git repository. It is the most widely deployed GitOps controller …

May 12·1 min read·79 words

GitOps

GitOps is an operational pattern in which the desired state of infrastructure and applications is declared in a Git repository, and an automated controller continuously reconciles the running system t…

May 12·1 min read·79 words

SLI

A Service Level Indicator (SLI) is a measurement of how well a service is performing on a given dimension, such as availability, latency, throughput, or correctness. SLIs are the raw observables; SLOs…

May 12·1 min read·82 words

SLO

A Service Level Objective (SLO) is an internal target for a service's reliability, expressed as a percentage over a window (for example, 99.9% of requests succeed within 300 ms over a rolling 28-day w…

May 12·1 min read·76 words

Grafana

Grafana is an open-source visualisation and dashboarding platform for time-series and observability data. It connects to many data sources (Prometheus, Loki, Tempo, Elasticsearch, MySQL, BigQuery, Clo…

May 12·1 min read·68 words

Prometheus

Prometheus is an open-source time-series database and monitoring system that scrapes metrics from instrumented applications over HTTP, stores them locally, and exposes a powerful query language (PromQ…

May 12·1 min read·72 words

OpenTelemetry

OpenTelemetry (OTel) is the CNCF standard for instrumenting applications to produce telemetry: traces, metrics, and logs. It defines an open specification, language SDKs, a vendor-neutral collector, a…

May 12·1 min read·72 words

Distributed Tracing

Distributed tracing captures the path of a single request as it travels through multiple services, recording the timing, parent-child relationships, and metadata of each step. A trace gives operators …

May 12·1 min read·84 words

Metrics

Metrics are numerical measurements aggregated over time, used for dashboards, alerting, and capacity planning. Where logs answer "what happened to this particular request", metrics answer "how is the …

May 12·1 min read·74 words

Logging

Logging is the practice of emitting timestamped records of events from running software so operators can reconstruct what happened. Modern systems use structured logging, where each record is a JSON o…

May 12·1 min read·71 words

Cache Stampede

A cache stampede (or thundering herd) is the failure mode where many concurrent requests for the same hot key all miss the cache at the same time, hit the origin simultaneously, and overwhelm it. Stam…

May 12·1 min read·87 words

TTL

TTL (Time To Live) is the duration for which a cached value, DNS record, or other transient piece of data remains valid. After the TTL expires, the entry is considered stale and either evicted, refres…

May 12·1 min read·83 words

Memcached

Memcached is an open-source, high-performance, distributed in-memory key-value cache, originally designed to speed up dynamic web applications by reducing database load. It is one of the oldest and mo…

May 12·1 min read·74 words

Eviction Policy

An eviction policy is the rule a cache uses to decide which entry to remove when it reaches its memory limit and needs to make room for a new entry. Choice of policy directly affects hit rate and is o…

May 12·1 min read·88 words

CDN

A Content Delivery Network (CDN) is a distributed network of edge servers that cache and serve content from locations close to the user, instead of every request reaching the origin server. CDNs reduc…

May 12·1 min read·85 words

Webhook

A webhook is an HTTP callback that one system sends to another to notify it of an event. Instead of the receiver polling for changes, the source POSTs an event payload to a URL the receiver registered…

May 12·1 min read·90 words

API Gateway

An API Gateway is a server that sits in front of a collection of backend services and exposes a single entry point to clients. It typically handles request routing, authentication, rate limiting, requ…

May 12·1 min read·75 words

Idempotency

Idempotency is the property that repeating an operation produces the same outcome as performing it once. An idempotent operation can be safely retried after a network failure, timeout, or partial fail…

May 12·1 min read·78 words

Pagination

Pagination is the API pattern for splitting a large collection into smaller pages so clients fetch results incrementally. It is one of the universal API design decisions, with consequences for perform…

May 12·1 min read·74 words

JSON

JSON (JavaScript Object Notation) is a lightweight, text-based data interchange format. It represents data as a small set of structures (objects, arrays, strings, numbers, booleans, null) that map nat…

May 12·1 min read·79 words

OpenAPI

OpenAPI (formerly Swagger) is a specification for describing HTTP APIs in a machine-readable format, typically YAML or JSON. An OpenAPI document defines an API's endpoints, request and response schema…

May 12·1 min read·73 words

HATEOAS

HATEOAS (Hypermedia As The Engine Of Application State) is the principle in Fielding's REST that responses must include links describing the next valid actions, so clients can navigate an application …

May 12·1 min read·81 words

gRPC

gRPC is a high-performance RPC framework developed at Google, using HTTP/2 as the transport and Protocol Buffers (protobuf) as the schema and serialization format. It is the default choice for interna…

May 12·1 min read·74 words

GraphQL

GraphQL is an API query language and runtime in which clients request exactly the fields they want from a typed schema. A single endpoint accepts queries, mutations, and subscriptions, and the server …

May 12·1 min read·77 words

HTTP

HTTP (Hypertext Transfer Protocol) is the application-layer protocol the web runs on. It defines how clients (typically browsers or API consumers) request resources from servers using methods like GET…

May 12·1 min read·80 words

IAM

Identity and Access Management (IAM) is the practice and tooling for defining who can do what on which resources within an organization or platform. It covers user and machine identities, group and ro…

May 12·1 min read·78 words

Authorization Server

An authorization server is the OAuth 2.0 component that authenticates the resource owner, obtains their consent, and issues access tokens (and optionally refresh tokens and ID tokens) to clients. It i…

May 12·1 min read·78 words

Bearer Token

A bearer token is a credential that grants the holder access to a resource simply by presenting the token, with no additional proof of identity. The defining property is that whoever holds the token m…

May 12·1 min read·85 words

JWE

JSON Web Encryption (JWE) is the encryption counterpart to JWS. Where JWS proves integrity and authenticity but leaves the payload readable, JWE encrypts the payload so only the intended recipient can…

May 12·1 min read·74 words

JWS

JSON Web Signature (JWS) is the cryptographic signing mechanism behind JWT. It defines how to produce and verify a signature over a JSON payload, using either symmetric (HMAC) or asymmetric (RSA, EC, …

May 12·1 min read·75 words

Refresh Token

A refresh token is a long-lived credential issued alongside a short-lived access token, used to obtain new access tokens without prompting the user to authenticate again. Refresh tokens trade convenie…

May 12·1 min read·82 words

Session

A session is the server-side state that represents an authenticated user across multiple requests. After a user signs in, the server creates a session record (storing the user ID, expiry, and any auxi…

May 12·1 min read·81 words

SAML

SAML (Security Assertion Markup Language) is an XML-based standard for exchanging authentication and authorization data between an identity provider (IdP) and a service provider (SP). It is the domina…

May 12·1 min read·78 words

PKCE

PKCE (Proof Key for Code Exchange), pronounced "pixy", is an extension to the OAuth 2.0 Authorization Code flow that protects against authorization code interception attacks. It was originally designe…

May 12·1 min read·77 words

OIDC (OpenID Connect)

OpenID Connect (OIDC) is an authentication layer built on top of OAuth 2.0. Where OAuth defines authorization (who can access what), OIDC defines authentication (who the user is) by adding a standardi…

May 12·1 min read·87 words

RAG

Retrieval-Augmented Generation (RAG) is a technique for letting a large language model answer using external documents instead of relying only on its training data. When a question arrives, the system…

May 12·1 min read·82 words

AI Agents

An AI agent is a system in which an LLM is given a goal and a set of tools, then loops through calling those tools, observing the results, and deciding what to do next, until the goal is reached or a …

May 12·1 min read·96 words

Context Window

The context window is the maximum amount of text an LLM can process in a single request, measured in tokens. It includes the system prompt, the user message, any retrieved context, prior conversation …

May 12·1 min read·87 words

Hallucination

Hallucination is the term for an LLM output that sounds confident but is unsupported by the provided context or factually wrong. It is the defining failure mode of generative models and the reason hig…

May 12·1 min read·76 words

Fine-tuning

Fine-tuning is the process of taking a pretrained model and continuing to train it on a target dataset, so its weights adapt to a specific task, domain, style, or output format. It is the standard way…

May 12·1 min read·85 words

Reranker

A reranker is a model that takes an initial set of retrieved candidates and re-orders them to improve precision. In a typical retrieval pipeline, a fast first-stage retriever returns 50 to 200 candida…

May 12·1 min read·83 words

Chunking

Chunking is the process of splitting documents into smaller passages before embedding them for retrieval. Chunk size and boundaries directly determine what a retrieval system can find: a chunk that is…

May 12·1 min read·78 words

Vector Database

A vector database is a storage system optimised for similarity search over high-dimensional vectors, typically embeddings. Given a query vector, it returns the closest vectors in the corpus using an a…

May 12·1 min read·72 words

Embeddings

Embeddings are dense vector representations of text, images, audio, or other data, learned so that semantically similar inputs end up near each other in vector space. They are the foundation of simila…

May 12·1 min read·76 words

React

React is an open-source JavaScript library for building user interfaces, created at Facebook (now Meta) and released in 2013. React introduced the component model and the virtual DOM that have since b…

May 12·1 min read·78 words

Nuxt

Nuxt is the production framework for Vue, analogous to Next.js for React. It provides file-based routing, server-side rendering, static site generation, layouts, auto-imports, a module ecosystem, and …

May 11·1 min read·69 words

Token Economics: What Every Developer Needs to Understand Now

The subscription model for AI was always a lie. GitHub just stopped pretending.

May 10·1 min read·22 words

Next.js

Next.js is the most widely used React framework, maintained by Vercel. It adds file-based routing, server rendering, static generation, incremental static regeneration, image and font optimisation, ed…

May 10·1 min read·71 words

Daily Hack: Sync Every Git Repo Before You Start Work

A daily bash hack I run every morning to fetch and sync all my git repos at once, before code reviews and standup.

May 9·7 min read·1291 words

JSX (JavaScript XML)

JSX (JavaScript XML) is a syntax extension to JavaScript that allows HTML-like markup to be written directly inside JavaScript code. JSX was popularised by React and is now widely used in other framew…

May 9·1 min read·75 words

Client-Side Rendering (CSR)

Client-Side Rendering (CSR) is the pattern where the server returns a minimal HTML shell and bundle of JavaScript, and the browser fetches data and builds the entire UI client-side. Single-page applic…

May 8·1 min read·82 words

Astro

Astro is a web framework built around the "islands architecture": pages are server-rendered HTML by default, with interactive components hydrated only where explicitly marked. Astro lets developers mi…

May 7·1 min read·73 words

Angular

Angular is an opinionated TypeScript front-end framework maintained by Google. Unlike React or Vue (libraries / progressive frameworks), Angular ships a full stack: routing, forms, HTTP client, depend…

May 6·1 min read·70 words

Spring Boot

Spring Boot is the most widely used Java framework for building production server applications. Built on top of the broader Spring Framework, Spring Boot adds auto-configuration, an embedded applicati…

May 5·1 min read·73 words

Rust

Rust is a systems programming language focused on memory safety without a garbage collector, achieved through a compile-time ownership model. It produces native binaries with C-like performance and ze…

May 4·1 min read·73 words

Ruby on Rails

Ruby on Rails (or just Rails) is a server-side web framework written in Ruby and released in 2004. Rails popularised the "convention over configuration" philosophy, the MVC pattern in web frameworks, …

May 3·1 min read·82 words

RabbitMQ

RabbitMQ is an open-source message broker that implements the AMQP protocol along with several others (MQTT, STOMP). Where Kafka is a log-oriented streaming platform, RabbitMQ is a flexible message ro…

May 2·1 min read·73 words

ORM

An ORM (Object-Relational Mapper) is a library that translates between application objects and relational database rows. ORMs handle query construction, parameter binding, result hydration into object…

May 1·1 min read·68 words

Node.js

Node.js is a JavaScript runtime built on the V8 engine that runs JavaScript outside the browser. It pioneered the use of a single-threaded, non-blocking, event-loop architecture for I/O-bound server w…

Apr 30·1 min read·79 words

Why You Should Not Build for 10 Million Users on Day One

Beautiful code. Clean architecture. Zero users. Most early-stage startups quietly die here, engineers solving problems three years too early while real users wait for something that simply works.

Apr 8·1 min read·40 words

Why You Should Not Build for 10 Million Users on Day One

Beautiful code. Clean architecture. Zero users. Most early-stage startups quietly die here, engineers solving problems three years too early while real users wait for something that simply works.

Apr 7·1 min read·40 words

Batch Processing in Modern Systems

Batching in distributed systems is like a tax strategy. Every database write, every API call, every network request carries overhead you can't avoid. The only question is how many times you pay it. Pr…

Apr 5·1 min read·55 words

MongoDB

MongoDB is an open-source document database that stores data as BSON documents grouped into collections. Each document is a self-contained JSON-like object with a flexible schema. Key features * F…

Mar 27·1 min read·71 words

Redis

Redis (Remote Dictionary Server) is an in-memory data store that holds its dataset in RAM for sub-millisecond access. Beyond simple key-value, Redis supports rich data structures including strings, ha…

Mar 27·1 min read·71 words

AWS

Amazon Web Services (AWS) is Amazon's cloud computing platform, comprising more than two hundred services for compute, storage, networking, databases, machine learning, and other capabilities, billed …

Mar 27·2 min read·245 words

Synchronized Expiration in Distributed Systems

Deterministic TTL in caching systems create hidden synchronization points that eventually break under scale. This deep dive explains how mature systems design for expiration, not just performance.

Mar 2·1 min read·32 words

An Ode to Stack Overflow: The Community That Taught a Generation to Think

Stack Overflow was not just a website. It was a place that taught us how to think in public. How to be precise. How to be wrong, and then better. We waited hours, sometimes days, for answers from stra…

Jan 15·1 min read·66 words

How Cybersecurity Will Evolve in 2026

Cybersecurity has collapsed under its own assumptions lately. Attackers scaled faster than humans could react, identity became the weakest link, and most defenses failed exactly when clarity mattered …

Jan 12·1 min read·48 words

Software Is Not Flexible. It Hardens as It Grows

Software has a reputation for being forgiving, something that is easy to change, easy to fix, and easy to reshape once we finally understand the problem we were supposed to be solving. That reputation…

Jan 5·6 min read·1190 words

Embedding Flutter Modules into Native Android and iOS Apps

We used Flutter Add-to-App to ship a side feature inside an existing Android and iOS app, saving weeks of duplicated work while keeping performance and user experience intact.

Dec 30·1 min read·37 words

MongoDB Data Modeling: How to Design Schemas for Real-World Applications

A fast MongoDB system comes from modeling data around how your application reads and writes it. This guide breaks down how to structure documents, when to embed or reference, the patterns used in real…

Dec 23·10 min read·1923 words

Your Career Isn’t a Ladder, It’s a Metamorphosis

Careers don’t move upward in straight lines. They reshape you through cycles of curiosity, chaos, stillness, and clarity. You aren’t climbing. You’re transforming.

Dec 20·4 min read·709 words

2026 Prediction: Foundation Models Are Becoming a Black Hole for AI Startups

Foundation models are no longer just platforms for AI startups. They are becoming gravity wells that pull entire product categories inward.. In 2026, AI startups built too close will be swallowed. The…

Dec 17·9 min read·1739 words

Beyond Paycheques: What Truly Keeps Employees Engaged

A candid HR perspective from Kriti Jain on why people stay: recognition, growth, flexibility, culture, and genuine human connection.

Dec 15·1 min read·26 words

The Strangler Fig Pattern: Growing Around Legacy

From the Leaning Tower of Pisa to the rainforest, the world shows us one truth: evolution beats replacement. The Strangler Fig pattern is the blueprint for evolving modern software without breaking wh…

Dec 13·6 min read·1163 words

Distributed Veto Power (and the Death of Speed)

A one-day feature turned into eight because too many people had the power to slow it down. Speed isn’t lost in engineering. It leaks through organizational cracks.

Dec 7·1 min read·35 words

Why Being Better Isn't Good Enough

Are you building a breakthrough business, or just a slightly better feature? Before you write a single line of code, use this two-step framework to validate your potential.

Dec 1·7 min read·1392 words

Why Rewrites Fail and Ugly Code Survives

Messiest parts of your codebase are usually the ones holding your entire business together. Rewriting them might be the fastest path to losing customers.

Nov 29·1 min read·31 words

Why the "Decentralized" Internet Keeps Breaking

The internet feels global and unbreakable, but it’s quietly centralized behind a handful of infrastructure giants. When one of them stumbles, the modern web forgets how to find itself.

Nov 24·1 min read·35 words

Think Outside the Product

Why did Zoom beat Skype? Why is CRED a status game, not a bill payment app?

Nov 11·1 min read·20 words

The Silent Rewrite of Software Engineering

The ground beneath software engineering is shifting, and the teams moving fastest aren’t rejecting LLMs, they’re weaponising them.

Nov 3·1 min read·24 words