Shashank Chakraborty — Backend Developer

AVAILABLE · BACKEND DEVELOPER @ VERTEX CLUB

FULL

BACKEND

DEV

I build systems that scale — graph-based security tools, real-time AI platforms, event-driven pipelines. Backend-first. Production-obsessed.

FastAPINeo4jDockerAWSPostgreSQLRedisCeleryReact

View Work Download CV

Scroll to explore01 / 05

shashank.config.ts

role: ""

focus: "Systems & AI"

stack: ["FastAPI", "Neo4j", "Docker"]

shipped: 3

latency: "sub-2s"

open: true

Latency Reduced

Users in Prod

Apps Shipped

RAG Accuracy

Dayananda Sagar CE · CGPA 8.38 · Bengaluru, IN

About

Engineer who ships.

I am a backend-focused developer who likes building systems with clear architecture, predictable performance, and production intent. My current work at Vertex Club anchors that mindset in a live platform.

I am especially drawn to AI tooling, graph databases, async workflows, and infrastructure that turns complexity into something operable.

I am looking for full-time backend or full-stack roles where I can own meaningful systems, ship fast, and keep improving the platform after launch.

Built production backend serving 200+ student members

Reduced API latency 40% via Redis caching + query optimization

Shipped 3 full-stack apps with RAG, graph traversal, and async pipelines

EXPERIENCE

Backend Developer

Vertex Club — Dayananda Sagar College of Engineering

2025 – Present

vertex.dsce.club

Built and shipped the production backend for vertex.dsce.club, serving 200+ student members and event registrations.

Designed RESTful APIs powering event registration workflows and the admin dashboard with Google OAuth via NextAuth.

Implemented role-based access control separating student and admin operations across a live platform.

Node.jsNext.jsPostgreSQLNextAuthVercel

NEO4JAWS IAMSECURITY

Shadow Permission Analyzer

Graph-based AWS IAM privilege escalation detector

Leverages BFS and DFS algorithms to identify hidden IAM permission chains that could lead to admin takeover. Extracts IAM data via boto3 and risk-scores each path by privilege level, chain depth, and escalation count.

Key DecisionChose Neo4j over a relational model because privilege escalation is fundamentally a graph traversal problem — SQL JOINs across self-referential tables would have been unreadable and slow.

Detects

Detects chains invisible to native AWS tooling

FastAPINeo4jReactboto3Python

View Code

User / Role

Risky Policy

Admin

GEMINI 2.0REAL-TIMEAI

SystemCraft

Real-time system design simulation platform

Stateful interview sessions with live AI evaluation powered by Gemini 2.0 Flash. Containerized with Docker, Nginx reverse proxy, and GitHub Actions CI/CD. MongoDB-backed persistent sessions with auto-save.

Key DecisionStreamed Gemini responses token-by-token rather than waiting for the full payload — this cut perceived latency from ~4s to under 2s without changing the model or infrastructure.

Sub-2s

Sub-2s structured response latency

Next.jsPythonMongoDBDockerGemini 2.0

Launch Platform View Code

SystemCraft — Live Session

LIVE

Architecture Canvas

Latency2000ms

Gemini Evaluation

REACTNODE.JSGEMINI AI

Interview Prep AI

Intelligent learning platform for interview preparation

Transforms interview preparation with dynamic, feedback-driven ecosystems. Features custom deck building from job descriptions, spaced repetition algorithms for long-term memory, and real-time voice practice with instant AI-powered critiques.

Key DecisionImplemented SM-2 spaced repetition algorithm server-side so review schedules persist across devices and sessions — not just a client-side timer.

Voice

Voice practice with instant AI feedback

ReactNode.jsExpressMongoDBGemini

Launch Platform View Code

Question

What is the time complexity of BFS?

Answer

O(V + E) — vertices plus edges

Voice Practice

OCR PIPELINEASYNC WORKERSDOCKER

DocuFlow

Intelligent invoice ingestion pipeline

Event-driven containerized pipeline with Celery workers, Redis queues, and PostgreSQL. Clean separation between OCR parsing and structured data extraction. Reduces document turnaround by 40% via async orchestration.

Key DecisionSeparated OCR parsing from structured extraction into two distinct Celery task queues — this let us scale the OCR workers independently during peak ingestion without touching the extraction layer.

40%

40% faster document processing

DockerCeleryRedisPostgreSQLPython

View Architecture GitHub

DocuFlow — Async Pipeline

847

Docs Processed

40%

Faster Turnaround

Pipeline Active

User / Role

Risky Policy

Admin

NEO4JAWS IAMSECURITY

Shadow Permission Analyzer

Detects

Detects chains invisible to native AWS tooling

FastAPINeo4jReactboto3Python

View Code

SystemCraft — Live Session

LIVE

Architecture Canvas

Latency2000ms

Gemini Evaluation

GEMINI 2.0REAL-TIMEAI

SystemCraft

Sub-2s

Sub-2s structured response latency

Next.jsPythonMongoDBDockerGemini 2.0

Launch Platform View Code

Question

What is the time complexity of BFS?

Answer

O(V + E) — vertices plus edges

Voice Practice

REACTNODE.JSGEMINI AI

Interview Prep AI

Voice

Voice practice with instant AI feedback

ReactNode.jsExpressMongoDBGemini

Launch Platform View Code

DocuFlow — Async Pipeline

847

Docs Processed

40%

Faster Turnaround

Pipeline Active

OCR PIPELINEASYNC WORKERSDOCKER

DocuFlow

40%

40% faster document processing

DockerCeleryRedisPostgreSQLPython

View Architecture GitHub

Skills

Backend-first toolkit.

BACKEND SYSTEMS

FastAPINode.jsExpressPythonREST APIsWebSockets

DATABASES & STORAGE

PostgreSQLMongoDBRedisNeo4jMySQLCelery

INFRASTRUCTURE & DEVOPS

DockerAWSGitHub ActionsNginxLinuxCI/CD

FRONTEND (WHEN NEEDED)

ReactNext.jsTypeScriptTailwind CSSFramer Motion

Let's build something.

Open to full-time backend and full-stack roles.

Available for new opportunities

GitHubgithub.com/Shashank0701-byte LinkedInlinkedin.com/in/shashank-chakraborty