Muhammad Habib Nur Aiman

BiteSight Analytics: Enterprise F&B Business Intelligence SaaS Live Access: https://bitesight.pages.dev/ BiteSight Analytics is an Enterprise-grade Business Intelligence (BI) SaaS tailored specifically for the Retail & F&B sectors. It merges premium visual aesthetics with advanced artificial intelligence to deliver an actionable, B2B analytical platform. 🧠 Core AI Technology: Agentic RAG & LLMs The brain of BiteSight is powered by an Agentic Retrieval-Augmented Generation (RAG) architecture designed for complex tabular data. Utilizing a tool-calling LangChain Agent integrated with the StepFun 3.5 Flash LLM (via OpenRouter), the system doesn't just read data—it reasons with it. The AI autonomously writes, executes, and iterates Python (Pandas) code within a secure backend sandbox to analyze thousands of CSV rows. It goes beyond simple data retrieval to generate hard Data Facts, deep Business Analyses, and Actionable Recommendations. 🚀 Key Features: • AI BizConsultant: A conversational interface allowing executives to "chat" directly with their metrics to uncover hidden operational bottlenecks. • Executive Dashboard V3: 11 highly interactive, premium visualizations (including Traffic Heatmaps, City-Level Treemaps, and Branch Comparison Radars). • Engineered Reporting: A seamless "One-Click PDF Export" utilizing an asynchronous DOM-to-SVG-to-PNG engine to stitch dashboards into presentation-ready A4 reports. • SaaS White-labeling: Precision theme customization featuring Glassmorphism UI and real-time Dark Mode transitions. 🛠️ Decoupled Serverless Architecture: • Frontend (UI/UX): Next.js 15+, Tailwind CSS, Recharts. Deployed as static HTML on Cloudflare Pages for 0ms edge latency. • Backend (AI & Processing): FastAPI, Python, Pandas, and LangChain. Containerized via Docker and deployed on Hugging Face Spaces as an isolated, secure execution sandbox. This project showcases end-to-end Full-Stack Data Science capabilities, bridging the gap between heavy AI computation and a sleek, client-facing SaaS experience.

RekrutPintar.id is a web-based recruitment platform designed to streamline the hiring process by moving beyond traditional CV screening. Utilizing the philosophy "CV's don't decide, RekrutPintar does," the system leverages AI to analyze candidate competency and behavioral data. It features an automated AI chat interview system that scores candidates in real-time, providing recruiters with a ranked dashboard of applicants based on potential and culture fit, drastically reducing administrative time.

Sistem klasifikasi citra aksara Jawa berbasis deep learning menggunakan ResNet18 yang di-fine-tune. Model dijalankan langsung di browser dengan ONNX Runtime Web dan menampilkan confidence score untuk setiap prediksi.

Saya menyediakan jasa pembuatan website UMKM profesional untuk usaha kuliner seperti kedai minuman dan makanan. Website dilengkapi dashboard admin sehingga pemilik usaha dapat mengelola menu, harga, dan ketersediaan produk secara mandiri tanpa perlu memahami teknis pemrograman. Website ringan, mobile-friendly, dan terintegrasi dengan pemesanan via WhatsApp.

Interactive dashboard for analyzing public sentiment related to Indonesia’s Ministry of Social Affairs (Kemensos). Includes sentiment distribution, time-based comment trends, and wordcloud + sample explorer for qualitative insights.

Browser-based AI system for detecting PPE compliance (helmet & safety vest) using YOLOv12. Runs fully client-side with ONNX Runtime Web, supporting both pretrained COCO and fine-tuned PPE models.