Bikash Talukder

Nexora AI — Personal Learning Platform

A futuristic AI learning platform combining chat, study mode, viva exams, and PDF export into one installable PWA — built for IELTS & BCS prep and beyond.

• AI Engine — Multi-key fallback (OpenRouter → Gemini Direct → Cloudflare AI Workers → Pollinations.ai) so it always has a working brain
• AI Compare Mode — Run the same question through multiple models side-by-side
• Online / Offline toggle — Built-in knowledge base + rule engine keeps it fully usable offline
• Study Mode — AI flashcards (text, image, or PDF → cards with flip animation + hints)
• Quiz — Multiple-choice with difficulty levels and AI re-explanation of wrong answers
• Spaced Repetition (SRS) — SM-2 algorithm with daily streaks and due badges
• Summarizer — Paste any text → bullets / paragraph / ELI5 / key terms / TL;DR, or compare two models
• Viva Mode — AI oral-exam simulator: question-by-question, scored critique, full grade report
• Progress Dashboard — Bar chart of study time, quiz accuracy sparkline, SRS progress, streak
• PDF Export — Print-ready SRS decks, quiz history stats table, Viva results with per-question feedback
• PWA — Installable on Android, iOS, desktop; offline service worker (cache-first assets, network-first app shell); Cloudflare AI Worker proxy for free, CORS-safe inference
• Voice — YouTube-style flashcard & summary playback · ElevenLabs voice mode · IELTS 4-part preparation · BCS preparation track

Healthcare Triage AI

An offline-first Progressive Web App for rural health workers — bilingual intake, voice input, and anomaly detection that runs without internet.

LiquiGuard — Liquidity Forecasting

SUST Final Hackathon winner-era project: EWMA-based liquidity forecasting with anomaly detection and live SSE streaming dashboards on PostgreSQL.

Cognexa AI

Zero-signup AI chat assistant in a monorepo — glass-morphic SPA front-end with local document extraction. Privacy-friendly by default.

Nexora Backend

Companion backend for the Nexora AI ecosystem — API surface that ties the PWA, study mode, and viva engine into a single deployable service.

bKash Complaint Parser

SUST Preliminary Hackathon microservice — a hybrid rule + LLM engine that parses and classifies bKash mobile-money complaints from raw text.

Cake E-Commerce Studio

An artisan cake studio storefront — modern React + TypeScript storefront with Tailwind styling, product catalog, and smooth checkout flow.

Coffee Shop E-Commerce

A viral-aesthetic coffee shop storefront built with Vite — AI-powered product discovery and edge-optimized RAG for fast recommendations.

Car Rental Management System

2nd-year Java project — a 3-layer Spring Boot 3 architecture for car rentals with H2 persistence, iText PDF invoices, Chart.js dashboards, and an AI assistant.

Blood Bank Management System

Full-stack Blood Bank platform with real-time dashboards, donor/recipient matching, and an operational v2 backend ready for production deployment.

Iskcon Youth Forum — Sylhet

A community/forum platform for the ISKCON youth chapter in Sylhet — built with TypeScript for fast iteration and clean moderation tooling.

Mobile Money Balance Viewer

A multi-provider balance dashboard (bKash, Nagad, Rocket) that unifies all your mobile-money wallets in one read-only TypeScript view.

ATM Simulation (Java 17)

A clean Java 17 ATM simulator — deposit, withdraw, balance, mini-statement, and PIN management with object-oriented design.

Mood Map AI

An AI mood-mapping TypeScript app that visualizes emotional patterns from text inputs into an explorable, color-graded map.

QueueStorm

A Python queue management system exploring worker scheduling, job retries, and backpressure handling for high-throughput pipelines.

Ticket Analyzer (SUST Hackathon)

SUST Hackathon submission — a support ticket analyzer that auto-categorizes, prioritizes, and summarizes incoming tickets in JavaScript.

MU Faculty Directory

A lightweight HTML directory of Metropolitan University faculty — departments, contacts, and office hours in one searchable page.

Realistic Map (C++)

A C++ graphics project rendering a realistic procedural map — terrain, water, and overlays built with native rendering primitives.