Level Up Coding

Git branching strategies: Do you know the differences? 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗯𝗿𝗮𝗻𝗰𝗵𝗶𝗻𝗴 keeps each feature in its own branch, isolated from the main branch; making pull requests easier to review. Once complete, the feature is merged back into main. 𝗚𝗶𝘁𝗳𝗹𝗼𝘄 uses two long-lived branches: dev for development and main for production. Features are built in separate branches, then merged into dev. Releases are prepared in dedicated branches before merging into production. 𝗚𝗶𝘁𝗟𝗮𝗯 𝗳𝗹𝗼𝘄 combines feature branching with environment-based deployment workflows. Changes move through environments like staging before reaching production, making it well-suited for CI/CD and staged releases. 𝗚𝗶𝘁𝗛𝘂𝗯 𝗳𝗹𝗼𝘄 simplifies things. The main branch is always deployable. Developers create short-lived branches, open pull requests, and merge once approved, often triggering deployment. 𝗧𝗿𝘂𝗻𝗸-𝗯𝗮𝘀𝗲𝗱 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 minimizes branching. Changes are merged into main frequently, supported by strong automated testing, CI pipelines, and feature flags to maintain stability. Each strategy solves the same core problem: How do teams move fast without breaking the system? But most issues don’t come from the branching model, they come from inside the branches: unclear changes, weak reviews, missing context. That’s what CodeRabbit Agent helps solve. A single agent inside your workflow that follows work end-to-end. It pulls your org’s context into one place and helps teams investigate, plan, and execute work directly from Slack. So instead of losing context between branches, your work stays connected as it evolves. 𝗧𝗿𝘆 𝗖𝗼𝗱𝗲𝗥𝗮𝗯𝗯𝗶𝘁 𝗔𝗴𝗲𝗻𝘁 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲 → lucode.co/coderabbit-age… What else would you add? —— ♻️ Repost to help others learn and grow. 🙏 Thanks to @coderabbitai for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at system design.

Git branching strategies: Do you know the differences? 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗯𝗿𝗮𝗻𝗰𝗵𝗶𝗻𝗴 keeps each feature in its own branch, isolated from the main branch; making pull requests easier to review. Once complete, the feature is merged back into main. 𝗚𝗶𝘁𝗳𝗹𝗼𝘄 uses two long-lived branches: dev for development and main for production. Features are built in separate branches, then merged into dev. Releases are prepared in dedicated branches before merging into production. 𝗚𝗶𝘁𝗟𝗮𝗯 𝗳𝗹𝗼𝘄 combines feature branching with environment-based deployment workflows. Changes move through environments like staging before reaching production, making it well-suited for CI/CD and staged releases. 𝗚𝗶𝘁𝗛𝘂𝗯 𝗳𝗹𝗼𝘄 simplifies things. The main branch is always deployable. Developers create short-lived branches, open pull requests, and merge once approved, often triggering deployment. 𝗧𝗿𝘂𝗻𝗸-𝗯𝗮𝘀𝗲𝗱 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 minimizes branching. Changes are merged into main frequently, supported by strong automated testing, CI pipelines, and feature flags to maintain stability. Each strategy solves the same core problem: How do teams move fast without breaking the system? But most issues don’t come from the branching model, they come from inside the branches: unclear changes, weak reviews, missing context. That’s what CodeRabbit Agent helps solve. A single agent inside your workflow that follows work end-to-end. It pulls your org’s context into one place and helps teams investigate, plan, and execute work directly from Slack. So instead of losing context between branches, your work stays connected as it evolves. 𝗧𝗿𝘆 𝗖𝗼𝗱𝗲𝗥𝗮𝗯𝗯𝗶𝘁 𝗔𝗴𝗲𝗻𝘁 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲 → lucode.co/coderabbit-age… What else would you add? —— ♻️ Repost to help others learn and grow. 🙏 Thanks to @coderabbitai for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at system design.

I wrote 35 articles for 35 system design concepts: 1) Microservices: lucode.co/microservices-… 2) Redis: lucode.co/redis-explaine… 3) Event-driven architecture: lucode.co/event-driven-a… 4) Database indexing: lucode.co/database-index… 5) Circuit breakers: lucode.co/circuit-breake… 6) ACID vs BASE: lucode.co/acid-vs-base-l… 7) Rate limiting: lucode.co/rate-limiting-… PS - if you want a structured path, get my free 142-page System Design Handbook when you join my free weekly newsletter → lucode.co/system-design-… 8) Observability: lucode.co/observability-… 9) System design quality attributes: lucode.co/system-design-… 10) CAP theorem: lucode.co/cap-theorem-li… 11) Idempotency: lucode.co/idempotency-in… 12) REST APIs: lnkd.in/gi5zd2jR 13) Sync vs Async: lnkd.in/gG-EnbA9 14) Network protocols: lucode.co/network-protoc… 15) Change Data Capture (CDC): lucode.co/change-data-ca… 16) CI/CD pipelines: lucode.co/ci-cd-lil1nlsm 17). SSO (single sign-on): lnkd.in/gXdVyqBg 18) JWT: lucode.co/json-web-token… 19) gRPC: lucode.co/grpc-explained… 20) Health checks vs heartbeats: lucode.co/health-checks-… 21) API gateway vs load balancer vs reverse proxy: lucode.co/api-gateway-vs… 22) HTTPS: lucode.co/https-explaine… 23) Load balancing algorithms: lucode.co/load-balancing… 24) Database caching: lucode.co/database-cachi… 25) API protocols: lucode.co/api-architectu… 26) CDN: lucode.co/cdn-lil1nlsm 27) Database types: lucode.co/database-types… 28) Message Queues: lucode.co/message-queues… 29) Hashing vs encryption vs tokenization: lnkd.in/gfqGYQnq 30) Service Discovery: lucode.co/service-discov… 31) Pub/Sub: lucode.co/pub-sub-lil1nl… 32) Connection pooling: lucode.co/connection-poo… 33) Forward proxy vs reverse proxy: lucode.co/forward-vs-rev… 34) Consistent hashing: lucode.co/consistent-has… 35) SQL vs NoSQL:lucode.co/sql-vs-nosql-l… —— 👋 PS: Get my free 142-page System Design Handbook when you join my free weekly newsletter. Join 33,000+ engineers → lucode.co/system-design-… —— ♻️ Repost to help others learn system design. ➕ Follow me ( Nikki Siapno ) to become good at system design.

What actually happens when you deploy to AWS? Deploying an app can feel like “just push code → it’s live.” In reality, your cloud is setting up compute, networking, and traffic routing before your app can even be reached. The simple mental model: Build → Provision → Run → Route → Shift traffic 𝟭) 𝗗𝗲𝗽𝗹𝗼𝘆 𝗶𝘀 𝘁𝗿𝗶𝗴𝗴𝗲𝗿𝗲𝗱 ↳ CI/CD pipeline starts and orchestrates the rollout. 𝟮) 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝗯𝘂𝗶𝗹𝘁 + 𝗽𝗮𝗰𝗸𝗮𝗴𝗲𝗱 ↳ Code is compiled, dependencies installed, artifact created. 𝟯) 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗶𝘀 𝗿𝗲𝘀𝗼𝗹𝘃𝗲𝗱 (𝗜𝗮𝗖) ↳ IaC defines the desired state, and AWS provisions resources to match it. 𝟰) 𝗖𝗼𝗺𝗽𝘂𝘁𝗲 𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀 𝗮𝗿𝗲 𝗰𝗿𝗲𝗮𝘁𝗲𝗱 ↳ VMs, containers, or functions are spun up. 𝟱) 𝗦𝘁𝗼𝗿𝗮𝗴𝗲 𝗶𝘀 𝗽𝗿𝗼𝘃𝗶𝘀𝗶𝗼𝗻𝗲𝗱 ↳ Databases, buckets, and volumes are created or attached. 𝟲) 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝘀𝘁𝗮𝗿𝘁𝘀 𝗿𝘂𝗻𝗻𝗶𝗻𝗴 ↳ Artifact is deployed and connects to services at runtime. 𝟳) 𝗧𝗿𝗮𝗳𝗳𝗶𝗰 𝗿𝗼𝘂𝘁𝗶𝗻𝗴 + 𝗛𝗧𝗧𝗣𝗦 𝗮𝗿𝗲 𝗰𝗼𝗻𝗳𝗶𝗴𝘂𝗿𝗲𝗱 ↳ Load balancer, DNS, and TLS are configured or reused. 𝟴) 𝗧𝗿𝗮𝗳𝗳𝗶𝗰 𝗶𝘀 𝘀𝗵𝗶𝗳𝘁𝗲𝗱 𝘁𝗼 𝘁𝗵𝗲 𝗻𝗲𝘄 𝘃𝗲𝗿𝘀𝗶𝗼𝗻 ↳ Rolling, blue-green, or canary deployment happens. The bigger question: Why does running code require all of this in the first place? Somewhere along the way, “deploying” turned into designing infrastructure, which feels like overkill when you just want to run something. Tools like exe[.]dev flip the flow: → Spin up a real machine instantly → Connect over SSH → It’s live You still get: → A persistent environment (it doesn’t disappear) → Automatic HTTPS (no networking setup) → Full control, without designing infra upfront It sits in a spot where it's more control than serverless, way less overhead than a VPS. If you're tired of infra decisions blocking momentum, it’s worth having on your radar. Check it out ➟ lucode.co/exe-dot-dev-z8… What else would you add? —— ♻️ Repost to help others learn and grow. 🙏 Thanks to @ssh_exe_dev for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at system design.

Traditional SDLC vs Modern SDLC vs Agentic SDLC 𝗧𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗦𝗗𝗟𝗖 (Waterfall era) is all about planning everything upfront and moving step-by-step through build and testing before shipping. It works when things are predictable, but becomes slow and costly the moment requirements change. 𝗠𝗼𝗱𝗲𝗿𝗻 𝗦𝗗𝗟𝗖 (Agile + DevOps) shifts to smaller iterations, continuous delivery, and fast feedback loops. It’s far more flexible, but humans are still responsible for most of the execution, including writing, testing, and fixing. 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗦𝗗𝗟𝗖 (emerging) introduces AI agents that can plan, code, test, and debug. Humans stay involved, but shift toward setting goals, reviewing outcomes, and guiding decisions, while agents handle more of the execution end-to-end. The shift isn’t just speed. It’s ownership of the work: Traditional → humans execute Modern → humans + automation Agentic → agents execute But here’s where most teams still break down: Even with AI, the SDLC is still fragmented. Planning happens in one tool. Code in another. PRs somewhere else. Context gets lost between every step. Agentic SDLC only works when that context doesn’t reset. That’s exactly what CodeRabbit Agent is built for. Instead of context resetting between tools, CodeRabbit Agent lives directly in your team’s workflow inside Slack, pulling in context from across your tools and systems and carrying it forward as work evolves. It can investigate issues, propose fixes, and open PRs, all within the same thread, carrying context forward as work moves from problem to resolution. If you want to see how agentic SDLC actually works in practice → lucode.co/coderabbit-age… What else would you add? —— ♻️ Repost to help others learn AI engineering. 🙏Thanks to @coderabbitai for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at AI engineering.

SLO vs SLI vs SLA An 𝗦𝗟𝗜 (Service Level Indicator) measures what’s actually happening in your system; like request latency, error rate, or uptime. It’s the raw signal that tells you how your service is performing. An 𝗦𝗟𝗢 (Service Level Objective) defines the acceptable range for an SLI; like “99.9% of requests succeed within 200ms.” It’s what your team aims to achieve to maintain reliability. An 𝗦𝗟𝗔 (Service Level Agreement) is a formal contract with users or customers, often tied to penalties if targets aren’t met. It defines the consequences, not just the goal. And when failures occur, that’s where postmortems should close the loop. But here’s the gap: Most postmortems are written after the fact. Digging through Slack. Rebuilding timelines. Guessing what actually happened. Which means they’re slow, and painful. So they get deprioritised, half-finished, or never written at all. That’s not just a process problem. It’s an experience and tooling problem. incident[.]io’s new postmortem workflow flips this: → It builds the draft for you from real incident data → So you start with context, not a blank page → It turns postmortems into a collaborative, structured workflow that’s actually easy to write and complete Worth a read → lucode.co/postmortems-re… SLIs tell you what happened. SLOs define what should happen. SLAs define what happens if you fail. Postmortems are where you make sure it doesn’t happen again. What else would you add? —— ♻️ Repost to help others learn and grow. 🙏 Thanks to @incident_io for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at system design.

12 GitHub repos to improve at AI engineering (categorized): 𝗟𝗟𝗠𝘀, 𝗥𝗔𝗚 & 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗦𝘆𝘀𝘁𝗲𝗺𝘀 1) Guides to build RAG, agents, vector search ↳ lucode.co/ai-developer-h… 2) RAG patterns ↳ github.com/NirDiamant/RAG… 3) E2E guide for building agents ↳ github.com/NirDiamant/age… 𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀 4) Prompt engineering ↳ github.com/dair-ai/Prompt… 5) Hands-on ML + deep learning ↳ github.com/ageron/handson… 6) Neural networks ↳ github.com/karpathy/nn-ze… 𝗖𝘂𝗿𝗮𝘁𝗲𝗱 𝗥𝗲𝘀𝗼𝘂𝗿𝗰𝗲 𝗟𝗶𝘀𝘁𝘀 7) Examples building with openAI ↳ github.com/openai/openai-… 8) Ready to run templates ↳ github.com/Shubhamsaboo/a… 9) ML frameworks, libs, & software ↳ github.com/josephmisiti/a… 10) Data Science resources ↳ github.com/academic/aweso… 𝗛𝗮𝗻𝗱𝘀-𝗢𝗻 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝘀 11) 100+ projects ↳ github.com/patchy631/ai-e… 12) E2E project with Claude Code ↳ github.com/shareAI-lab/le… If you're serious about AI engineering, this repo is worth bookmarking (star it): lucode.co/ai-developer-h… It’s not just theory; you’ll find: • Full AI apps and reference implementations • Jupyter notebooks for RAG, agents, and vector search • Practical guides on agent architecture and memory systems If you found this list useful, star the repo so you can come back to it later (and support more content like this) What other AI GitHub repos should be on the list? —— ♻️ Repost to help others learn AI engineering. 🙏 Thanks to @Oracle for sponsoring this post. ➕ Follow me ( Nikki Siapno ) to improve at AI engineering.

REST vs GraphQL vs gRPC (explained in under 2 mins): 𝗥𝗘𝗦𝗧 is resource-based and uses standard HTTP methods like GET, POST, PUT, and DELETE. It’s simple, widely adopted, and easy to debug, but can lead to over-fetching or under-fetching when clients need specific data. 𝗚𝗿𝗮𝗽𝗵𝗤𝗟 lets clients request exactly the data they need through a single endpoint. It reduces over-fetching and improves flexibility, but adds complexity around schema design, caching, and performance tuning. 𝗴𝗥𝗣𝗖 is a high-performance RPC framework that uses Protocol Buffers and HTTP/2. It’s efficient, strongly typed, and ideal for internal service-to-service communication, but less human-readable and harder to use from browsers. Where they typically fit: REST is often used for public APIs, GraphQL for complex frontend data needs, and gRPC for internal service communication. Each one shifts complexity to a different place. You’re not choosing an API, you’re choosing tradeoffs. Full breakdown (with visuals) here → lucode.co/api-protocols-… What else would you add? ♻️ Repost to help others learn system design. ➕ Follow me ( Nikki Siapno ) to improve at system design.

We’ve invested deeply in security at Replit, including our recent launches with Security Agent + Auto-Protect. If you want to move your app to Replit, we’re offering free app imports for a limited time. Bring your app over and keep building safely. Get started here: replit.com/free-import