Bản xem trước sẽ hiển thị ở đây sau khi tạo xong.
Trợ lý AI
Create a portfolio website with the following details:
Theme: Modern Dark
Pages: Home, Projects, Expertise, About, Contact
Texts: --- Home Page ---
Purpose: To provide an immediate "Who, What, and Why" within 5 seconds.
Hero Section: * Headline: [Name] | Fullstack Architect & AI Implementation Expert
Sub-headline: "Building scalable web applications from the first line of code to AI-powered production."
Call to Action (CTA): [View My Work] or [Let's Talk].
The "Elevator Pitch": A 2-sentence summary of your 5 years in IT.
Tech Stack Ribbon: A scrolling or static bar of logos (e.g., React, Node.js, Python, AWS, OpenAI API, TensorFlow).
Featured Project: A single, high-impact card showing your best AI-integrated application.
--- Projects Page ---
Purpose: To prove you can handle the entire lifecycle of a product.
Project 1: SupplyChain AI – Predictive Inventory Engine
The Problem: A mid-sized retail client was losing 15% of annual revenue due to overstocking perishable goods and frequent "out-of-stock" scenarios on high-demand items.
The Solution: I built an end-to-end forecasting platform from scratch. The system ingests historical sales data via a custom ETL pipeline and uses a Temporal Fusion Transformer (TFT) model to predict stock needs with 92% accuracy.
The AI Edge: Unlike static thresholds, this AI accounts for local weather patterns and social media trends to adjust stock levels dynamically.
Tech Stack: Next.js, FastAPI, PostgreSQL, AWS SageMaker, and Docker.
[From Scratch Badge]: Designed the database schema, trained the model, and configured the CI/CD pipeline for automated deployments.
Project 2: MediScan: Intelligent Patient Triage Portal
The Problem: High-volume clinics struggled with long wait times because administrative staff couldn't prioritize patients based on the urgency of their symptoms described in intake forms.
The Solution: A secure, HIPAA-compliant web portal that uses Natural Language Processing (NLP) to analyze patient-submitted symptoms in real-time. It flags high-risk keywords and "Red Flag" symptoms, moving urgent cases to the top of the doctor’s dashboard.
The AI Edge: Implemented a custom-tuned BERT model to categorize medical intent and sentiment, ensuring that "chest pain" is prioritized over "mild cough" without human intervention.
Tech Stack: React, Express.js, MongoDB, Python (HuggingFace), and Azure Health Bot.
[From Scratch Badge]: Managed everything from the encrypted database architecture to the frontend UI and the SSL/security hardening.
Project 3: SentinelLog – Autonomous DevOps Monitor
The Problem: A DevOps team was overwhelmed by "alert fatigue," spending hours digging through thousands of system logs to find the root cause of server crashes.
The Solution: I developed a fullstack monitoring tool that uses Anomaly Detection (Isolation Forests) to identify irregular patterns in system logs before a crash occurs. It visualizes the "health score" of a cluster in a real-time dashboard.
The AI Edge: The system learns the "baseline" behavior of your specific servers and only triggers alerts when a pattern deviates significantly from the norm, reducing false positives by 70%.
Tech Stack: Go (Golang), Vue.js, Redis, TimescaleDB, and TensorFlow.js.
[From Scratch Badge]: Built the data ingestion engine in Go for high performance and deployed the entire stack via Kubernetes (K8s).
Project 4: LegalEase – Semantic Contract Auditor
The Problem: Legal teams were manually reviewing 50+ page contracts to ensure compliance with new regional data privacy laws, a process taking 4–6 hours per document.
The Solution: A web-based "Auditor" application where users drop a PDF and get an instant AI-generated report highlighting non-compliant clauses. I implemented a RAG (Retrieval-Augmented Generation) architecture to "teach" the AI the latest legal statutes.
The AI Edge: Uses vector embeddings to compare contract text against a live database of legislation, providing a "compliance score" and suggested rewrites for risky paragraphs.
Tech Stack: TypeScript, NestJS, Pinecone (Vector DB), LangChain, and OpenAI API.
[From Scratch Badge]: Architected the PDF-to-Text parsing engine, the vector search logic, and the responsive frontend interface.
--- Expertise Page ---
Purpose: To explain how you apply AI to real-life use cases—this is your unique selling point.
Service Block 1: Fullstack Engineering
"I build monolithic or microservice architectures that scale. My focus is on clean, maintainable code and seamless UX."
Service Block 2: AI Implementation
"I move AI beyond the chatbot. I specialize in integrating LLMs, predictive modeling, and automated data processing into existing business workflows."
Service Block 3: End-to-End Deployment
"From CI/CD pipelines to cloud infrastructure management. Your app stays live and stays fast."
--- About Page ---
Purpose: To build trust and show the person behind the code.
Professional Timeline: A brief visual journey of your 5 years in the industry.
The Philosophy: Why you focus on AI. (e.g., "I believe AI shouldn't just be a gimmick; it should be a functional layer that saves users time.")
Personal Touch: One sentence about what you do when you aren't coding (e.g., "When I'm not fine-tuning models, I'm usually [hiking/gaming/cooking].")
--- Contact Page ---
Purpose: To make it as easy as possible for a recruiter or client to reach you.
Contact Form: Fields for [Name], [Email], and [Message].
Direct Links: * LinkedIn: [Link]
GitHub: [Link]
Email: [Your Professional Email]
Availability Status: A small green dot saying "Currently accepting new projects/roles."
Images: