Shubham Upadhyay

Senior Backend Architect

Open to Collaborate
Bengaluru, IN
Connect Channels
Architecture Teardowns & Deep Dives

Battle-Tested Systems.
Deep-Dive Case Studies.

Every production system is a series of trade-offs: latency vs. memory, simplicity vs. flexibility, or consistency vs. throughput. Here I break down the actual engineering decisions behind my work—the constraints encountered, alternatives evaluated, why option A was rejected over option B, and how the architecture holds up in production.

You'll find end-to-end platform teardowns alongside specialized deep dives on lightweight Go daemons, WebAssembly execution sandboxes, and zero-lock database optimization pipelines.

14
Total Studies
6
System Architectures
8
Technical Deep Dives
ArchMCP Teardown
Architecture

Zero-dependency in-memory graph index resolving cross-service dependencies in <1.5ms without Neo4j.

Node.js MCP Protocol O(1) Lookup
ClipNest Linux Daemon
Deep Dive

Skipping Electron for a 14MB native Go background daemon with Wayland clipboard hooks and local SQLite.

Go SQLite Privacy Toggle
Browser PHP WASM
Client Runtime

Compiling PHP 8.2 into WebAssembly to execute arbitrary user scripts in-browser with zero server execution.

WebAssembly Web Workers $0 Server Cost
Showing 14 of 14 case studies

Displaying all production case studies, including full system teardowns and specialized engineering decision articles.

System Architecture WEB APPLICATION
8 min read Dec 2021 – Sep 2022

Letus Code (E-commerce ERP) — Architecture & System Teardown

Manufacturing to order platform with CRM, order tracking and automation.

Letus Code is a custom manufacturing and quotation platform where clients can request CNC machining parts, get instant quotes, and track order progress. The platform integrates with Shopify and provides a complete workflow from inquiry to dispatch.

System Architecture DEVELOPER TOOL / AI
8 min read 2024 – Present
Project: ArchMCP

ArchMCP — Architecture & System Teardown

Giving AI coding assistants a live map of your microservices, APIs, and dependencies.

ArchMCP scans polyglot microservices to build an in-memory architecture index. Instead of pasting entire repositories into prompts, AI coding assistants can query only the specific services, endpoints, and downstream dependencies they need.

Engineering Deep Dive System Architecture & DX
6 min read Aug 2024
Project: ArchMCP

Why We Skipped Neo4j & Postgres: Building a Zero-Dependency, In-Memory Architecture Index

How a deliberate refusal to over-engineer made ArchMCP fast, zero-dependency, and instantly runnable in under 3 seconds.

Why microservice architecture graphs do not need heavy database clusters, and how building an in-memory graph index with thread-safe locks and automated secret masking created a far better developer experience.

Engineering Deep Dive Distributed Systems
7 min read Sep 2024
Project: ArchMCP

Before You Break Production: Mapping Cross-Service Dependencies in Microservices

Why grep misses breaking changes in distributed systems, and how ArchMCP links HTTP calls, Kafka topics, background jobs, and Docker configs to warn you before you deploy.

How we built an asynchronous dependency engine that traces direct and transitive ripples across polyglot microservices, classifying risk severity before PRs are merged.

Engineering Deep Dive AI & Context Engineering
6 min read Oct 2024
Project: ArchMCP

Why Dumping 20 Repositories Into an AI Fails (And What We Built Instead)

How moving from passive prompt stuffing to active on-demand MCP tool retrieval turned 50,000-token hallucination traps into 1,500-token precision context packages.

Why copying dozens of microservices into AI coding assistants degrades model reasoning, and how designing surgical MCP tools like find_api_owner and get_database_schema keeps Claude, Cursor, and Antigravity fast, accurate, and cost-effective.

System Architecture DEVELOPER TOOL / CLI
8 min read 2024 – Present
Project: SQL Doctor

SQL Doctor — Architecture & System Teardown

A developer-first terminal toolkit to inspect, diagnose, optimize, and safely manage SQL databases.

SQL Doctor is a lightweight, single-binary terminal toolkit written in Go to diagnose slow queries, inspect schema health, compare databases, and validate migration safety before production rollout. Built around deterministic database facts first, with optional Gemini AI assistance.

System Architecture DEVELOPER PLATFORM
8 min read 2023 – Present
Project: ShubhamLabs

ShubhamLabs — Architecture & System Teardown

Building a Developer Platform from Real Engineering Problems

ShubhamLabs is a developer platform I built to bring together developer tools, practical engineering notes, architecture experiments, and open-source projects rooted in real software engineering problems.

Engineering Deep Dive DevOps Tooling
5 min read Feb 2024
Project: ShubhamLabs

Building a Dockerfile & Docker Compose Generator

How a frustrating, repetitive setup habit turned into an instant, 100% client-side container boilerplate engine.

Turning project requirements into ready-to-use Dockerfile and docker-compose.yml files directly in the browser — zero backend, zero boilerplate lookups.

System Architecture SYSTEM UTILITY / GO
8 min read 2023 – Present
Project: ClipNest

ClipNest — Architecture & System Teardown

A lightweight, privacy-first clipboard daemon and glassmorphism web manager.

ClipNest is a lightweight clipboard manager for modern Linux desktops (Wayland) built in Go. It runs as a quiet systemd user daemon, stores history in a local SQLite file, includes an instant pause toggle for privacy, and serves a modern local web dashboard.

Engineering Deep Dive Go & Linux Systems
5 min read Dec 2023
Project: ClipNest

Why I Skipped Electron: Building a 14MB Native Clipboard Manager in Go

How pairing a quiet Wayland background daemon with an embedded local web UI replaced a 400MB memory hog.

Why modern clipboard tools shouldn't consume 400MB of RAM or log your passwords, and how building a native Go daemon for Linux Wayland with embedded SQLite and a local web UI delivered a fast, privacy-first desktop utility.

System Architecture ENTERPRISE MIS / WORKFLOW
8 min read 2018 – 2021

InfoTech MIS & Work Manager — Architecture & System Teardown

Internal enterprise operational backbone and project task delivery platform for GIS, LiDAR, and engineering drafting.

InfoTech MIS served as the operational nervous system for an engineering enterprise specializing in GIS, Underfloor Heating, LiDAR, and Photogrammetry. It unified internal administrative processes (daily attendance regularizations, leave approvals, automated payslips) with a high-throughput production work manager that automated project creation, file dispatching to engineers, active task timers, multi-stage QA verification, and direct-to-client server publishing.

Engineering Collaboration

Architecting a high-concurrency backend or complex system?

I design resilient database architectures, build low-latency APIs in Go and PHP, and craft developer tools designed to handle production pressure with zero bloat.

Chat on WhatsApp
Navigating...