+FULL-STACK DEVELOPMENT • COMPUTER VISION • WEB PLATFORMS
Full-Stack Developer • Computer Science Student

Rudra Pratap Singh

Building responsive web applications, role-based business platforms, and applied computer vision systems. Currently pursuing B.Tech in Computer Science & Business Systems at Sister Nivedita University, Kolkata.

CIIR 2026
Second Author, Q-Sentry Paper
Semantic-Quantum Ethical AI Model
1st Place
Hackfest 2025 Winner
Sister Nivedita University
Handshake AI
Freelance Software Engineer
Frontend Architecture & Features
2024–2028
B.Tech Undergrad (CSBS)
Sister Nivedita University, Kolkata
EDITORIAL PERSPECTIVE

A Perspective on
Architecture & Utility

PHILOSOPHY • ARCHITECTURAL PRINCIPLES
“I do not view engineering as merely assembling generic templates. Every system is a synthesis of architectural discipline, data integrity, and tactile human interaction.”

Whether fine-tuning an RT-DETR transformer for fine-grained vehicle localization or deploying multi-tenant RBAC permissions for enterprise clients, the objective is constant: durable software that solves real problems with clarity and verifiable performance.

ISYSTEM COHESION

Full-Stack System Cohesion

Architecting clean, type-safe workflows from PostgreSQL relational models and MongoDB aggregations to high-throughput REST API handlers and React 19 server components.

IIDEEP LEARNING

Applied Computer Vision & AI

Building verifiable machine learning pipelines—from RT-DETR detection transformers and PyTorch classification models to high-speed LLM assistants using the Groq SDK.

IIICRAFTSMANSHIP

Editorial Interaction Design

Crafting responsive interfaces with architectural grids, accessible design tokens (WCAG AA), smooth Lenis inertia, and tactile micro-interactions that feel immediate.

IVPRAGMATISM

Real-World Operational Delivery

Shipping practical tools for real businesses—from enterprise field tracking for Excel Technologies to voice-assisted MSME store accounting with MunshiJi AI.

ACADEMIC FOUNDATION • SISTER NIVEDITA UNIVERSITY

Bachelor of Technology in Computer Science & Business Systems (CSBS)

Sister Nivedita University, Kolkata (2024–2028). Synthesizing core computer science algorithms, deep learning vision models, and distributed systems with enterprise business workflow analysis and product strategy.

Secondary & Senior Secondary Foundation:Kolkata, WB • Science (PCM & Computer Science)
Research Acceptance
Q-Sentry: CIIR 2026 (Second Author)
Semantic-Quantum Ensemble for Ethical Classification
International Delegation
IoT Student Exchange • Sichuan, China
Sichuan Vocational Institute of Technology (2026)
Competitive Distinction
Hackfest 2025 Regional & State Winner
TECHNICAL DISCIPLINES

Architectural Capabilities& Verified Tooling

FIVE SPECIALIZED DOMAINS
SCENE: 01 / 05
DOMAIN 01 • Vision & Intelligence

AI & Computer Vision

Deep learning models, computer vision pipelines, and agentic LLM orchestration

Execution Pipeline
01Image Ingestion [Camera / File]
02→ RT-DETR Detection Transformer [Localization]
03→ EfficientNet PyTorch Classifier [Attribute Inference]
04→ HTML5 Konva Canvas [Interactive Visual Bounding Boxes]
Implementation Evidence
Source Excerpt
// RT-DETR Detection Transformer & PyTorch Inference Pipeline (Computer Vision)
import torch
from transformers import RTDetrForObjectDetection, RTDetrImageProcessor

processor = RTDetrImageProcessor.from_pretrained("PekingU/rtdetr_r50vd")
model = RTDetrForObjectDetection.from_pretrained("PekingU/rtdetr_r50vd")

def detect_and_classify_vehicles(image_tensor):
    inputs = processor(images=image_tensor, return_tensors="pt")
    with torch.no_grad():
        outputs = model(**inputs)
    results = processor.post_process_object_detection(
        outputs, target_sizes=[image_tensor.size[::-1]], threshold=0.75
    )
    return results
Production VerifiedType Safe
Verified Tooling:
PyTorch[Production]
RT-DETR[Computer Vision]
EfficientNet[Classification]
Groq AI SDK[LLM Orchestration]
RAG Systems[Retrieval]
DOMAIN 02 • Frontend Architecture

Frontend Architecture

Component systems, responsive editorial layouts, and high-performance client state

Execution Pipeline
01Server Components (RSC) / Vite SPA
02→ TanStack Query [Async Cache & Deduplication]
03→ Zod Schema Validation [Form State Integrity]
04→ Tailwind v4 + Radix UI [Accessible Design System]
Implementation Evidence
Source Excerpt
// React 19 Async Action & Optimistic State Pipeline (Frontend Architecture)
"use client";
import { useOptimistic, useTransition } from "react";
import { updateClientRecordAction } from "@/app/actions/crm";

export function ClientWorkflowEditor({ client, onSync }) {
  const [isPending, startTransition] = useTransition();
  const [optimisticState, setOptimisticState] = useOptimistic(
    client,
    (state, update) => ({ ...state, ...update, syncStatus: "optimistic" })
  );

  const handleUpdate = (formData) => {
    startTransition(async () => {
      setOptimisticState(formData);
      await updateClientRecordAction(formData);
    });
  };
}
Production VerifiedType Safe
Verified Tooling:
React 19[Modern Patterns]
Next.js 15 (App Router)[Production]
TypeScript[Strict Typing]
Tailwind CSS v4[Design Tokens]
TanStack Query[Async State]
DOMAIN 03 • Backend & Systems

Backend & Systems

REST API design, serverless microservices, security protocols, and state synchronization

Execution Pipeline
01Client Request [Bearer / HTTP-only Cookie]
02→ Express 5 / Vercel Serverless Function
03→ JWT Verification & Team-Scoped RBAC Middleware
04→ MongoDB Atlas Mongoose / PostgreSQL Transactions
Implementation Evidence
Source Excerpt
// Express 5 JWT Middleware & Role-Based Isolation (Backend Systems)
import jwt from "jsonwebtoken";

export const requireTeamAuth = (allowedRoles = []) => {
  return async (req, res, next) => {
    const token = req.cookies.auth_token;
    if (!token) return res.status(401).json({ error: "Unauthorized session" });

    const decoded = jwt.verify(token, process.env.JWT_SECRET);
    if (allowedRoles.length && !allowedRoles.includes(decoded.role)) {
      return res.status(403).json({ error: "Insufficient privilege" });
    }
    req.user = decoded;
    next();
  };
};
Production VerifiedType Safe
Verified Tooling:
Node.js[Runtime]
Express 5[Server Architecture]
Next.js API Handlers[Serverless]
JWT Cookie Auth[Security]
Zod[Schema Validation]
DOMAIN 04 • Databases & Cloud

Cloud, Database & Tooling

Data modeling, multi-tenant databases, cloud storage, and automated deployment

Execution Pipeline
01Application Queries
02→ Prisma ORM / Mongoose ODM
03→ PostgreSQL / MongoDB Atlas Managed Cluster
04→ Cloudinary Signed Media & AWS S3 Object Storage
Implementation Evidence
Source Excerpt
// Prisma Data Access & PostgreSQL Transaction Isolation
datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
}

model OrganizationEmployee {
  id           String   @id @default(uuid())
  email        String   @unique
  role         Role     @default(EMPLOYEE)
  shifts       Shift[]
  leaveAudits  Leave[]
  createdAt    DateTime @default(now())
}
Production VerifiedType Safe
Verified Tooling:
MongoDB Atlas[NoSQL Modeling]
PostgreSQL[Relational Data]
Prisma ORM[Data Access]
Supabase Realtime[WebSockets]
Cloudinary[Media Pipeline]
DOMAIN 05 • Creative Development

Creative Development & Interaction

WebGL shaders, kinetic typography, smooth physics motion, and responsive micro-interactions

Execution Pipeline
01User Input [Pointer & Scroll]
02→ Lenis Inertia Engine + GSAP ScrollTrigger
03→ Vanilla Three.js WebGL Spatial Scene [Pavilion Datum & Camera Lerp]
04→ React 19 View Transitions [Hardware-Accelerated 60fps]
Implementation Evidence
Source Excerpt
// Three.js Halftone Relief Sculptural System
import * as THREE from "three";

const scene = new THREE.Scene();
const circleGeometry = new THREE.CircleGeometry(0.038, 12);
const alabasterMaterial = new THREE.MeshStandardMaterial({
  color: 0xF2F0EA,
  roughness: 0.35,
  metalness: 0.05,
});
const instancedMesh = new THREE.InstancedMesh(circleGeometry, alabasterMaterial, 12000);
scene.add(instancedMesh);
Production VerifiedType Safe
Verified Tooling:
Vanilla Three.js[WebGL Graphics]
GSAP & ScrollTrigger[Motion Choreography]
Lenis[Smooth Scroll]
Kinetic Typography[Character Physics]
Framer Motion[Micro-interactions]
SCROLL VERTICALLY TO TRAVERSE CAPABILITY DOMAINS
SELECTED WORK

Ideas are a start.This is the work.

BUILT, DEPLOYED & TESTED

Curated case studies covering full-stack enterprise systems, deep learning computer vision pipelines, and production web applications.

EtechCRM•See case study
01 / 06•
Client ManagementField Visit TrackingTanStack Query & ZodVercel Serverless

Client Management & Field Visit Tracking Platform

Enterprise Client: Excel Technologies & Services

A production client management and field operations tracking platform built for Excel Technologies & Services, architected as a modular Single Page Application backed by Node.js serverless functions.

Excel Technologies • Field Operations & Client CRM
PRODUCTION DEPLOYMENT
Active Field Visits
18 / 20
90% Target Met
Enterprise Accounts
142
Multi-Tenant Scoped
GPS Verification
100%
Zero Coordinate Drift
Sync Engine
TanStack
Optimistic Cache
REAL-TIME FIELD VISIT AUDIT STREAMKOLKATA METROPOLITAN REGION
Technocraft Engineering LtdAnnual Maintenance
22.5726° N, 88.3639° E•Exec: A. Banerjee
VERIFIED & SYNCED14:32 IST
Eastern Power Systems CorpTransformer Audit
22.5851° N, 88.4124° E•Exec: S. Mukherjee
IN TRANSIT15:10 IST
Metro Rail Logistics HubField Equipment Check
22.5697° N, 88.3562° E•Exec: R. Sen
SCHEDULED16:30 IST
TECH STACK • VERIFIED LOGOS & PROTOCOLS
React 19
TypeScript
Vite
exExpress 5
MongoDB Atlas
Cloudinary
MunshiJi AI•See case study
02 / 06•
Voice-Assisted OperationsAutomated GST InvoicingCredit Ledger TrackingInventory Management

Intelligent Retail Billing & Store Ledger Platform

An AI-powered billing, inventory management, and credit accounting platform designed to streamline store operations for retail businesses and shop owners through voice assistance and automated invoicing.

MunshiJi AI • MSME Retail Billing & Voice Ledger Engine
VOICE & LEDGER SYNCED
Natural Voice Ingestion StreamGROQ LPU • 42ms
“Record 25 units heavy brass valve and 12 PVC elbows for Gupta Hardware on 30-day credit ledger”
Target Account:Gupta Hardware (Credit Client)
Line Items:2 SKU Items Parsed • 100% Match
Automatic OCR IngestionMultilingual Hindi/English
Digital GST Tax InvoiceINV-2025-0842
Heavy Brass Valve (25 @ ₹450)₹11,250.00
PVC Industrial Elbows (12 @ ₹250)₹3,000.00
CGST (9%) + SGST (9%)₹2,565.00
Total Ledger Debit
₹16,815.00
ESC/POS Status
80mm Print Dispatched
TECH STACK • VERIFIED LOGOS & PROTOCOLS
Next.js App Router
TypeScript
Node.js
exExpress
MongoDB
Groq API
CardecaAlgo•See case study
03 / 06•
RT-DETR Detection TransformerPyTorch EfficientNetInteractive Konva CanvasExportable Inspection Reports

RT-DETR & EfficientNet Vehicle Intelligence

A research-driven computer vision platform pairing Real-Time Detection Transformers (RT-DETR) with a custom PyTorch EfficientNet classifier for real-time vehicle localization and fine-grained classification.

CardecaAlgo • RT-DETR Vision Transformer & PyTorch Inference
60 FPS INFERENCE • 16.4ms
SUV • 98.7%
[124, 76, 492, 340] • IOU: 0.94
SEDAN • 96.2%
[510, 110, 720, 320] • EfficientNet
STREAM: 1080p60RESNET-50 BACKBONE
HTML5 KONVA CANVAS
LATENCY
16.4ms / frame
DETECTION HEAD
RT-DETR Transformer
CLASSIFIER
PyTorch EfficientNet
EXPORT FORMAT
ONNX & TensorRT
TECH STACK • VERIFIED LOGOS & PROTOCOLS
Next.js App Router
TypeScript
PyTorch
RT-DETR
EfficientNet-B4
Konva
ETHOS•See case study
04 / 06•
Groq AI SDK AssistantPrisma ORMMulti-Tier RBACFullCalendar Scheduling

AI-Assisted Human Resource Management System

A comprehensive Human Resource Management System featuring an integrated Groq AI assistant, multi-tenant role-based access control, attendance tracking, and interactive scheduling.

Ethos • Enterprise HRMS & AI Knowledge Copilot
ORGANIZATION ACTIVE
Employee Presence & Shift Ledger98% Synced
Engineering Lead (Remote)Clocked In • 09:12 IST
Product Manager (HQ)Clocked In • 09:30 IST
UI/UX Designer (On Leave)Statutory Leave Approved
Enterprise AI Policy AssistantRAG Engine
“What is the statutory annual and sick leave entitlement for full-time engineering staff?”
> Retrieved from 2025_HR_Handbook.pdf: 18 annual days + 12 medical leaves with full rollover.
TECH STACK • VERIFIED LOGOS & PROTOCOLS
Next.js App Router
TypeScript
@ai-sdk/groq
Prisma ORM
PostgreSQL
FullCalendar
RelancePay•See case study
05 / 06•
Multi-Currency & GSTDynamic PDF GenerationJWT Cookie AuthResend Email Integration

Freelance Multi-Currency Invoicing SaaS

A production-grade invoicing SaaS engineered for independent freelancers and consultants to handle domestic compliance (INR with GST) alongside international multi-currency client billing (USD).

RelancePay • Multi-Currency Fintech Gateway & Ledger
IDEMPOTENCY LOCKED
USD SETTLEMENT
$24,850.00
Webhook Verified
EUR FX CONVERSION
€18,220.00
Real-time Spot Rate
INR DOMESTIC ROUTE
₹12,45,000
IMPS / NEFT Dual Route
TECH STACK • VERIFIED LOGOS & PROTOCOLS
Next.js App Router
TypeScript
Tailwind CSS
shadcn/ui
MongoDB Atlas
Mongoose
VAYRIN•See case study
06 / 06•
Emergency Dispatch EngineSupabase Real-TimeTrauma Center RoutingHealthcare Coordination

Emergency Healthcare Coordination Platform Concept

An architectural platform concept and system specification designed to coordinate multi-node emergency healthcare dispatch between ambulances, emergency call centers, and hospital trauma units.

Vayrin • Emergency Healthcare Routing & Geospatial Triage
CODE RED DISPATCH
Critical Ambulance Dispatch • Sector V Trauma Unit
ETA: 6.4 mins • Hospital ICU Bed Reserved • Dijkstra Geospatial Graph
PRIORITY 1 DISPATCH
TECH STACK • VERIFIED LOGOS & PROTOCOLS
Next.js
TypeScript
Supabase Realtime
PostgreSQL
Tailwind CSS v4
CAREER CHRONOLOGY

Engineering Roles
& Academic Foundation

PROFESSIONAL ROLES • 2024–2028
HIGHER EDUCATION • 2024–2028

Bachelor of Technology in Computer Science & Business Systems (B.Tech CSBS)

Sister Nivedita University, Kolkata (2024–2028). Comprehensive curriculum bridging core computing algorithms, computer vision pipelines, and distributed databases with corporate technology management and enterprise business strategy.

Institution
Sister Nivedita University, Kolkata
Tenure • Status
2024–2028 (Undergraduate CSBS)
HIGHER SECONDARY (CLASS XII)2022–2024

Senior Secondary • Science (PCM & CS)

Physics, Chemistry, Mathematics, and Computer Science. Rigorous academic foundation in mathematical reasoning, advanced calculus, and algorithm design.

Kolkata, WB • Pure Science Stream
SECONDARY EDUCATION (CLASS X)Completed 2022

Secondary School Board Examination

Comprehensive curriculum across sciences, advanced mathematics, social sciences, and computer applications. Graduated with academic excellence.

Kolkata, WB • Foundational Sciences
Jul 2026 — Present
Handshake AI
Freelance Software Engineer
Intelligent systems, full-stack application development, AI integrations
Core Contributions & Verified Delivery:
Architected and deployed responsive full-stack features and AI integrations.
Engineered scalable backend service handlers and optimized client state pipelines.
Developed robust API integrations ensuring secure data handling and high responsiveness.
Mar 2025 — Jun 2025
GeekOTee
Development Intern
Frontend architecture, responsive component engineering, state management
Sep 2024 — Feb 2025
InternStudio
Software Engineering Intern
Web development, UI/UX implementation, cross-platform responsiveness
RESEARCH & HONORS

Peer-Reviewed Research
& Competitive Honors

PEER-REVIEWED PUBLICATION • CIIR 2026
PEER-REVIEWED PUBLICATION • CIIR 2026

Q-Sentry: A Semantic-Quantum Ensemble for Ethical Classification

Peer-reviewed research publication accepted at CIIR 2026, investigating ethical content classification using semantic embeddings and quantum ensemble principles.

Author Role: Second Author•Conference: CIIR 2026•Domain: Ethical Classification & Quantum Machine Learning
ACCEPTANCE STATUS
Peer-Reviewed Paper Accepted, CIIR 2026 (Second Author)
PRIMARY FOCUS
Semantic-Quantum Ensemble Modeling
Verified Competitive Honors & International Programs:
Hackfest State Championship • 2025
1st Place Winner

Hackfest 2025

Secured first place across both regional and state rounds for full-stack architecture and real-time execution.

Google Developer Groups • 2025
Runner-Up

GDG APL Kolkata

Awarded runner-up honor for developing Clean Room RAG—a dual-mode research intelligence platform that indexes academic literature (arXiv) and private documents using local LLMs.

OpenAI & Outskill • 2025
National Finalist

OpenAI x Outskill AI Builders

Selected as a national finalist for building full-stack AI-native product prototypes and agentic workflows.

Level Supermind • 2025
Finalist

Level Supermind Hackathon

Earned finalist standing for building real-time data analysis workflows and automated insights dashboards under strict hackathon evaluation criteria.

Sichuan Vocational Institute for Technology, China • 2026
International Delegate

International IoT Student Exchange Program

Selected as an academic delegate for the International IoT Student Exchange Program at Sichuan Vocational Institute for Technology in China (Sep–Oct 2026), focusing on embedded IoT protocols, hardware interfacing, and distributed systems.