Shivam Verma

Backend Engineer

Shivam Kumar Verma

API Architecture · Distributed Systems

I design and build scalable backend systems, REST & event-driven APIs, and data pipelines. Focused on performance, observability, and clean architecture.

View ProjectsView Resume

Experience

Internships and professional roles.

Contributing

·VerlyAI

Contributing to the VerlyAI project

Jan 2026 – Present

  • ▸Solely architected the end-to-end AI chatbot flow, building a robust data ingestion pipeline with a custom web crawler to scrape, process, and index website content from raw URLs.
  • ▸Developed the core AI response service from the ground up, optimizing the backend architecture for ultra-low latency and highly accurate customer query resolution.
  • ▸Designed and implemented a seamless human-escalation system for routing unresolved queries to live agents, alongside an automated lead-generation module to drive customer acquisition.
  • ▸Built a comprehensive WhatsApp integration from scratch enabling true omnichannel support, and integrated a secure end-to-end payment processing system for user subscriptions.
TypeScriptNext.jsNode.jsPostgreSQLGolang

Full Stack Developer Intern

·Klimb·(Remote)

Oct 2025 – Jan 2026

  • ▸Migrated 1,000+ legacy subscriptions to a modular credit-based architecture, ensuring 100% data integrity through pre-migration impact analysis with zero downtime.
  • ▸Engineered a metered billing system with real-time quota enforcement, resolving financial leaks and improving cost-tracking accuracy by 80% for AI and email services.
  • ▸Developed an AI resume parsing pipeline to automate metadata extraction, processing 1000+ resumes daily and reducing manual screening bottlenecks by 100%.
  • ▸Automated sourcing workflows by synchronizing application-phase metadata, saving an estimated 40 hours per week by eliminating redundant data entry across all modules.
  • ▸Built a centralized Super Admin module to manage 1000+ global tiers and credit allocations, streamlining operational control and reducing administrative overhead by 90%.
TypeScriptReactNode.jsExpress.jsMongoDBRedis

Engineering Timeline

Journey through key projects and milestones.

2024

CodeSM – Role-based Coding Platform

Built a coding platform with problem storage on AWS and isolated Docker execution.

Node.jsReactMongoDBAWSDocker

Rate limiting and container isolation are critical for secure code execution.

2024

DeepDoc – RAG Document Chat

AI-powered document Q&A using embeddings and vector search with Gemini.

Next.jsLangChainGeminiPinecone

Chunking strategy and embedding model choice directly impact retrieval quality.

2024

Chatterly – Real-time Chat

Group chat with WebSockets, Redis caching, and PostgreSQL persistence.

Node.jsReactPostgreSQLRedisWebSockets

Redis for session/cache and PostgreSQL for source of truth keeps latency low.

2024

Trading-n8n – Workflow Automation

n8n-based workflows for trading logic and integrations with TypeScript/JavaScript.

TypeScriptn8nJavaScript

Workflow automation enables repeatable pipelines and clear step separation.

2024

Email-rag – Email Thread RAG

Chat over email threads with FastAPI, BM25 retrieval, and Gemini with message-level citations.

PythonFastAPIReactGeminiDocker

RAG over structured threads with citations improves trust and debuggability.

Featured Backend Projects

Engineering case studies: problem, architecture, and learnings.

CodeSM

RepoLive

Problem

Need a secure, scalable platform for storing coding problems and running user code in isolation.

Architecture

Frontend (React) → API (Node.js) → Problem storage (AWS S3) → Code execution (Docker containers). Rate limiting at API layer.

Node.jsReactMongoDBAWSDocker

DeepDoc

RepoLive

Problem

Users need to query large PDFs via natural language without reading entire documents.

Architecture

Upload → Chunking (LangChain) → Embeddings (Gemini) → Pinecone. Chat API uses retrieval + LLM to answer.

Next.jsLangChainGeminiPineconeVercel Blob

Chatterly

RepoLive

Problem

Real-time group chat with persistence and low latency at scale.

Architecture

React client ↔ WebSocket server (Node) ↔ Redis (cache/session) + PostgreSQL (persistence).

Node.jsReactPostgreSQLRedisWebSockets

Trading-n8n

Repo

Problem

Automate trading and workflow pipelines using n8n with custom logic and integrations.

Architecture

n8n workflows with TypeScript/JavaScript nodes for trading logic and external API integrations.

TypeScriptn8nJavaScriptCSS

Email-rag

Repo

Problem

Chat over email threads with answers grounded in the actual messages and citations.

Architecture

React UI + FastAPI backend. BM25 retrieval (thread-scoped or global), Gemini for answers with message-level citations.

PythonFastAPIReactGeminiDockerBM25

Competitive Programming

Algorithm practice and contest journey across platforms.

Problem-solving focus: Data structures, algorithms, dynamic programming, and contest participation. Track record of 700+ problems across platforms.

Technical Stack

Backend, data, infrastructure, and core concepts.

Backend

  • Node.js
  • Python
  • FastAPI
  • Express
  • REST APIs
  • WebSockets

Databases

  • PostgreSQL
  • MongoDB
  • Redis
  • Vector Databases
  • SQLAlchemy

Infrastructure

  • Docker
  • Linux
  • Git
  • CI/CD
  • AWS
  • Render

Concepts

  • System Design
  • Caching
  • Authentication
  • Database Indexing
  • API Rate Limiting
  • Event-driven Architecture

Engineering Blog

Technical articles on backend systems and APIs.

View all posts

Certificates

Credentials with verification links. Replace verifyUrl with your actual links.

AWS Cloud Practitioner

Amazon Web Services

Verify

MongoDB Basics

MongoDB University

Verify

Node.js Backend Development

Coursera / Similar

Verify

System Design Fundamentals

Educative / Similar

Verify

GitHub Activity

Contribution graph, pinned repos, and top languages.

Contribution graph

Embed your GitHub contribution graph here (e.g. via activity graph or image API).

Top languages

JavaScript45%
Python25%
TypeScript20%
Other10%

Pinned repositories

CodeSM

Role-based coding platform

DeepDoc

RAG document chat

Chatterly

Real-time chat

Trading-n8n

n8n trading workflows

Email-rag

Email thread RAG with FastAPI

View full profile on GitHub

Get in Touch

Open to backend engineering roles and collaboration. Reach out via any channel below.

EmailGitHubLinkedInTwitter