Freelancer logo
Log In Sign Up
Post a Project

EXPLORE

Amazon Web Services Chatbot Machine Learning (ML) Python Web Hosting
Profile cover photo
User Avatar

Yu X.

@xuezhe315

monthly-level-three.svg
4.9
7

4.6

4.6

100%

🏆Top 5% AI & Software Developer

$40 USD / Hour

・

Flag of

China (12:29 AM)

・

Joined on May 21, 2025

・

1 Recommendation

$40 USD / Hour

・

I am a seasoned software engineer with a proven track record of delivering cutting-edge, user-centric digital solutions for startups, enterprises, and global brands. My work spans the entire software development lifecycle — from ideation and architecture to deployment and optimization — with a focus on modern web, mobile, and AI-driven applications. Over the years, I’ve mastered the art of blending aesthetics, performance, and scalability, ensuring every project I touch not only meets technical requirements but also drives measurable business growth. 🔹 My Core Expertise Web Development (Frontend & Backend) Modern JavaScript frameworks: React.js, Vue.js, Next.js Backend mastery: Node.js, Django, FastAPI, Laravel Styling & UI: Tailwind CSS, Material UI, custom component libraries REST & GraphQL APIs, real-time applications, PWA development Mobile Development Cross-platform excellence: React Native, Expo High-performance native integrations Pixel-perfect, responsive mobile UI/UX Artificial Intelligence & Machine Learning Natural Language Processing (NLP), computer vision, predictive modeling Custom model training & deployment (TensorFlow, PyTorch, scikit-learn) GPT, Vertex AI, OpenAI API integrations for chatbots, automation & intelligent decision-making Data pipelines, analytics dashboards, AI-powered automation Cloud & DevOps AWS, GCP, Azure deployment & scaling Docker, Kubernetes, CI/CD pipelines, serverless architectures PostgreSQL, MySQL, MongoDB, Firebase 💼 Why Clients Choose Me Proven Delivery: Successfully led 50+ full-scale projects from concept to production. Business-Driven Mindset: I translate technical complexity into real-world business outcomes. Scalable Solutions: Architected systems that support millions of users without compromising performance. Transparent Communication: Clear, proactive, and client-focused at every stage. Whether you’re launching a high-growth SaaS platform, building an AI-powered mobile app, or modernizing your digital ecosystem, I bring a unique blend of deep technical expertise and strategic thinking to turn complex visions into flawless reality.

Verifications

On time

97%

On budget

97%

Accept rate

100%

Repeat hire rate

13%

Freelancer Python Developers China
See all Python Developers in China

Portfolio

Portfolio

13580508
I designed and implemented a Retrieval-Augmented Generation (RAG) based support automation system that leverages historical Zendesk ticket data to improve customer service efficiency. The system fetches and processes resolved tickets, transforming them into a structured knowledge base that serves as the foundation for intelligent response generation. The workflow begins with automated data ingestion from Zendesk, where resolved tickets are retrieved and cleaned. These records are then indexed into a vector database, enabling semantic search and contextual retrieval. When new support tickets arrive, the system performs similarity searches across the knowledge base to extract the most relevant past resolutions. The results are passed into a large language model (LLM) that generates accurate, context-aware responses tailored to the query. To support ongoing improvement and usability, I implemented: Interactive testing mode, allowing support teams to validate AI-suggested responses befo

RAG-based Support Agent

13578565
I designed and developed a fully automated WhatsApp chatbot using Python and Twilio’s WhatsApp Business API. The bot serves as an interactive portfolio assistant, allowing potential clients to explore my services, projects, and skills directly through WhatsApp in a conversational manner. The bot was built to: Present portfolio work dynamically: Users can view project categories, navigate through samples, and request details with simple text commands. Answer FAQs automatically: Handles common client questions about services, pricing models, and working process. Streamline communication: Instead of browsing external websites, clients can access my offerings via a familiar messaging platform. Enable lead capture: The bot collects user details (name, email, project requirements) and stores them securely for follow-ups. Provide 24/7 availability: Ensures potential clients get instant responses even outside working hours. What makes this project special is its fusion of automation and

WhatsApp chatbot

13564672
I developed a cutting-edge food ordering app that leverages AI-powered chatbots to enhance customer engagement and streamline the ordering process. The app allows users to browse menus, customize their orders, and place them easily through an intuitive interface. The AI chatbot acts as a virtual assistant, guiding customers through the entire ordering process with personalized recommendations based on their preferences, past orders, and real-time promotions. The AI chatbot uses natural language processing (NLP) to understand customer queries, suggest dishes, and even handle complaints or order modifications in a conversational manner. It also offers voice recognition, allowing customers to place orders by simply speaking to the bot, making the experience even more seamless. The backend of the app is built with Node.js and MongoDB, ensuring scalability and performance, while the front-end is crafted using React Native for cross-platform support. I implemented real-time order tracking,

Food Ordering App with AI Chatbot

13564593
I developed a robust and efficient delivery app solution that empowers businesses to streamline their order processing, delivery tracking, and customer service. The app supports both Android and iOS platforms, ensuring accessibility for a wide range of users. It features real-time GPS tracking for customers and drivers, dynamic route optimization, and integrated payment processing for seamless transactions. The app allows customers to place orders, track deliveries in real-time, and rate their experience, while drivers can receive task notifications, view optimized routes, and complete deliveries with ease. The back-end system is built using modern technologies like Node.js and PostgreSQL for high performance and scalability. For mobile development, I used React Native, which ensures a fast and native-like experience across both platforms. The integration of Firebase ensures that notifications, user authentication, and data storage are seamless and efficient. What sets this project a

Delivery App With Real-Time Tracking

13122840
OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

13122796
This project showcases the development of an advanced AI-driven solution designed to streamline and enhance the pre-construction planning and cost estimation process in the AEC (Architecture, Engineering, and Construction) industry. Leveraging the power of Vertex AI models (Google Cloud), the system performs a comprehensive range of tasks including: ? 1. Narrative & Design Specification Generation Automatically generates detailed narrative descriptions for construction projects. Supports dynamic input of desired building attributes such as area, number of floors, occupancy type, and project location. Produces tailored design plans and schematic outlines using natural language inputs. ? 2. AI-Driven Cost Estimation Breaks down and estimates labor cost, material cost, and equipment cost based on standard databases and regional pricing trends. Supports customizable cost factors and inflation adjustments. Outputs professional-grade Excel and PDF reports for client proposals or i

AI Assistant Construction Estimation

This project showcases the development of an advanced AI-driven solution designed to streamline and enhance the pre-construction planning and cost estimation process in the AEC (Architecture, Engineering, and Construction) industry. Leveraging the power of Vertex AI models (Google Cloud), the system performs a comprehensive range of tasks including: ? 1. Narrative & Design Specification Generation Automatically generates detailed narrative descriptions for construction projects. Supports dynamic input of desired building attributes such as area, number of floors, occupancy type, and project location. Produces tailored design plans and schematic outlines using natural language inputs. ? 2. AI-Driven Cost Estimation Breaks down and estimates labor cost, material cost, and equipment cost based on standard databases and regional pricing trends. Supports customizable cost factors and inflation adjustments. Outputs professional-grade Excel and PDF reports for client proposals or i

AI Assistant Construction Estimation

13580508
I designed and implemented a Retrieval-Augmented Generation (RAG) based support automation system that leverages historical Zendesk ticket data to improve customer service efficiency. The system fetches and processes resolved tickets, transforming them into a structured knowledge base that serves as the foundation for intelligent response generation. The workflow begins with automated data ingestion from Zendesk, where resolved tickets are retrieved and cleaned. These records are then indexed into a vector database, enabling semantic search and contextual retrieval. When new support tickets arrive, the system performs similarity searches across the knowledge base to extract the most relevant past resolutions. The results are passed into a large language model (LLM) that generates accurate, context-aware responses tailored to the query. To support ongoing improvement and usability, I implemented: Interactive testing mode, allowing support teams to validate AI-suggested responses befo

RAG-based Support Agent

13578565
I designed and developed a fully automated WhatsApp chatbot using Python and Twilio’s WhatsApp Business API. The bot serves as an interactive portfolio assistant, allowing potential clients to explore my services, projects, and skills directly through WhatsApp in a conversational manner. The bot was built to: Present portfolio work dynamically: Users can view project categories, navigate through samples, and request details with simple text commands. Answer FAQs automatically: Handles common client questions about services, pricing models, and working process. Streamline communication: Instead of browsing external websites, clients can access my offerings via a familiar messaging platform. Enable lead capture: The bot collects user details (name, email, project requirements) and stores them securely for follow-ups. Provide 24/7 availability: Ensures potential clients get instant responses even outside working hours. What makes this project special is its fusion of automation and

WhatsApp chatbot

13564672
I developed a cutting-edge food ordering app that leverages AI-powered chatbots to enhance customer engagement and streamline the ordering process. The app allows users to browse menus, customize their orders, and place them easily through an intuitive interface. The AI chatbot acts as a virtual assistant, guiding customers through the entire ordering process with personalized recommendations based on their preferences, past orders, and real-time promotions. The AI chatbot uses natural language processing (NLP) to understand customer queries, suggest dishes, and even handle complaints or order modifications in a conversational manner. It also offers voice recognition, allowing customers to place orders by simply speaking to the bot, making the experience even more seamless. The backend of the app is built with Node.js and MongoDB, ensuring scalability and performance, while the front-end is crafted using React Native for cross-platform support. I implemented real-time order tracking,

Food Ordering App with AI Chatbot

13564593
I developed a robust and efficient delivery app solution that empowers businesses to streamline their order processing, delivery tracking, and customer service. The app supports both Android and iOS platforms, ensuring accessibility for a wide range of users. It features real-time GPS tracking for customers and drivers, dynamic route optimization, and integrated payment processing for seamless transactions. The app allows customers to place orders, track deliveries in real-time, and rate their experience, while drivers can receive task notifications, view optimized routes, and complete deliveries with ease. The back-end system is built using modern technologies like Node.js and PostgreSQL for high performance and scalability. For mobile development, I used React Native, which ensures a fast and native-like experience across both platforms. The integration of Firebase ensures that notifications, user authentication, and data storage are seamless and efficient. What sets this project a

Delivery App With Real-Time Tracking

13122840
OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

OVERVIEW This app helps industrial operators and engineers: • Ask natural language questions about their plant/process • Train the system using real events from SCADA or historian time-series data • Validate and use the system in production for anomaly detection, diagnostics, and optimization ________________________________________ ✅ PHASE 1: DOCUMENT Q&A (Knowledge Activation via RAG) Purpose Unlock value from existing domain knowledge before time-series data is used. SME knowledge becomes searchable and usable immediately. How It Works 1. Upload documents (PDF, Word, TXT, CSV) 2. System chunks, embeds, and indexes content (RAG pipeline) 3. User asks a natural language question 4. System retrieves relevant content and generates an answer 5. Sources are shown for trust and traceability ? Common Document Types • Control Narratives • SOPs (Standard Operating Procedures) • Alarm Response Manuals • Cause & Effect Matrices • Tag Dictionaries (with Tag ID descriptions) • Operating L

Document Q&A + Time-Series Learning

13122796
This project showcases the development of an advanced AI-driven solution designed to streamline and enhance the pre-construction planning and cost estimation process in the AEC (Architecture, Engineering, and Construction) industry. Leveraging the power of Vertex AI models (Google Cloud), the system performs a comprehensive range of tasks including: ? 1. Narrative & Design Specification Generation Automatically generates detailed narrative descriptions for construction projects. Supports dynamic input of desired building attributes such as area, number of floors, occupancy type, and project location. Produces tailored design plans and schematic outlines using natural language inputs. ? 2. AI-Driven Cost Estimation Breaks down and estimates labor cost, material cost, and equipment cost based on standard databases and regional pricing trends. Supports customizable cost factors and inflation adjustments. Outputs professional-grade Excel and PDF reports for client proposals or i

AI Assistant Construction Estimation

This project showcases the development of an advanced AI-driven solution designed to streamline and enhance the pre-construction planning and cost estimation process in the AEC (Architecture, Engineering, and Construction) industry. Leveraging the power of Vertex AI models (Google Cloud), the system performs a comprehensive range of tasks including: ? 1. Narrative & Design Specification Generation Automatically generates detailed narrative descriptions for construction projects. Supports dynamic input of desired building attributes such as area, number of floors, occupancy type, and project location. Produces tailored design plans and schematic outlines using natural language inputs. ? 2. AI-Driven Cost Estimation Breaks down and estimates labor cost, material cost, and equipment cost based on standard databases and regional pricing trends. Supports customizable cost factors and inflation adjustments. Outputs professional-grade Excel and PDF reports for client proposals or i

AI Assistant Construction Estimation

4.9 · 7 Reviews
Reviews
D

Fix RAG Chatbot Tokens • $100 USD

Knows his stuff and delivered an outstanding outcome on time.

Datonyx5

@Datonyx5

•

Flag of

Melbourne, Australia

•

1 month ago

Share

P

Python AI RESTful Web App -- 2 • $750 USD

Good job. Intelligent and expert in many fields

Phani V.

@phanivedula

•

Flag of

Conroe, United States

•

2 months ago

Share

User Avatar

Install RWHOIS service on Debian 12 server • $70 USD

1 - The project was not delivered on time. 2 - I was informed that the project had been completed, but the documentation was poor and full of errors. 3 - The service provider wanted more money to complete what my company had already paid for. 4 - It took too long to respond. Poor quality service and no professional commitment! I do not recommend hiring this professional.

Netfire T.

@netfiretec

•

Flag of

Montes Claros, Brazil

•

4 months ago

Share

C

ML Speech Analysis & Dashboard Upgrade • $900 USD

Extremely great worker with exceptional understanding of the project's goals and objectives. I can't wait to work long term with a very skilled person!

Leon T.

@contactlishan

•

Flag of

Glasgow, United Kingdom

•

5 months ago

Share

User Avatar

Project for Yu X. set a plan and diagrame for workfolw with comfyui • $10 USD

To be honest, this guy talked to me in the chat about the project, he was familiar with all the details, unfortunately we didn't agree because the project was bigger than my budget, and in appreciation of his interest and time I am writing these words and I am ready to deal with him again because I sensed seriousness and professionalism from him.

Mohammed T.

@mohdkhld1234

•

Flag of

Jeddah, Saudi Arabia

•

6 months ago

Share

Experience

Senior Full Stack Developer

Jan, 2022 - Apr, 2025

•

3 years, 2 months

Shenyang CloudTech Solutions

Jan, 2022 - Apr, 2025

•

3 years, 2 months

Led development of a financial compliance portal for a government client using Node.js, Express, and React. Built RESTful APIs for loan approval workflows, integrated third-party APIs (credit checks, payment providers), and designed real-time reporting dashboards. Optimized PostgreSQL queries and MongoDB structures for high-volume transactions. Tech stack: Node.js, Express, React, MongoDB, PostgreSQL, AWS, GraphQL, API Gateway.

Flag of

Shenyang, China

Jan, 2022 - Apr, 2025

•

3 years, 2 months

Senior Back End Developer

Jan, 2017 - Dec, 2021

•

4 years, 11 months

Liaoning Digital Systems

Jan, 2017 - Dec, 2021

•

4 years, 11 months

Architected backend services for loan processing and eligibility verification. Implemented real-time payment tracking using Node.js, PostgreSQL, and AWS serverless. Integrated payment gateways (Stripe, PayPal APIs) and built secure configuration options. Delivered daily reports and analytics dashboards for portfolio monitoring. Mentored junior developers on secure coding and compliance-ready API design. Tech stack: Node.js, TypeScript, Nest.js, Express, GraphQL, PostgreSQL, REST APIs.

Flag of

Shenyang, China

Jan, 2017 - Dec, 2021

•

4 years, 11 months

Mid-Level Back End Developer

Oct, 2014 - Dec, 2016

•

2 years, 2 months

Shenyang FinTech Innovations

Oct, 2014 - Dec, 2016

•

2 years, 2 months

Developed custom fintech-style apps for retail/finance clients, integrating with Shopify and payment providers. Built GraphQL and REST APIs for transaction processing. Implemented fraud detection pipelines (Kafka + Redis + anomaly models). Automated repayment and delinquency alerts with cron-based workflows. Tech stack: Node.js, PHP (Laravel), MongoDB, SQL, Redis, Vercel, Prisma, GraphQL.

Flag of

Shenyang, China

Oct, 2014 - Dec, 2016

•

2 years, 2 months

Education

Shenyang University

2006 - 2010

•

4 years

Bachelor's Degree in Computer Science

Flag of

China

2006 - 2010

•

4 years

Freelancer logo US (International) / English
Help & Support

Freelancer

Categories Projects Contests Freelancers Enterprise Innovation Challenges AI Development Membership Preferred Freelancer Program Project Management Local Jobs Photo Anywhere Showcase API for Developers Get Verified

About

About us How it Works Security Investor Sitemap Stories News Team Awards Press Releases Careers

Terms

Privacy Policy Terms and Conditions Copyright Policy Code of Conduct Fees and Charges

Partners

Escrow.com Loadshift Warrior Forum

Apps

Apple App Store logoGoogle Play logo

Registered Users

Total Jobs Posted

Freelancer ® is a registered Trademark of Freelancer Technology Pty Limited (ACN 142 189 759)

Copyright © 2025 Freelancer Technology Pty Limited (ACN 142 189 759)