We are seeking an experienced developer to create plugins for CorelDRAW. The ideal candidate should have a strong understanding of CorelDRAW's architecture and plugin development process. You will be responsible for designing, coding, and testing plugins to enhance functionalities. This role requires creativity and technical skills to develop solutions that meet our needs. If you have a passion for graphic design tools and a knack for programming, we would love to hear from you!

# Upwork Job Post: LiveKit + LangGraph Conversation Flow Expert Needed

## Job Title
**Experienced Python Developer for LiveKit + LangGraph Conversation Flow Integration**

## Project Overview
We're seeking an experienced Python developer to enhance our AI voice agent system that combines LiveKit Agents with LangGraph for dynamic conversation flows. We have existing code that needs refinement and optimization to work seamlessly with LiveKit's voice pipeline.

## What We Have
- Complete conversation flow schema (Pydantic models)
- Basic ConversationFlowGraph implementation using LangGraph StateGraph
- LiveKit Agents integration setup
- Working voice pipeline foundation

## What We Need
- Optimize the conversation flow graph to work perfectly with LiveKit's voice synthesis
- Implement proper node transitions and edge conditions
- Ensure the LLM responses integrate correctly with LiveKit's streaming architecture
- Debug and resolve any voice output issues
- Add robust error handling and logging

## Technical Requirements

### Must Have Experience With:
- **LiveKit Agents** framework
- **LangGraph** for building state machines
- **LangChain** integration with LiveKit
- **Pydantic** for data validation
- **Voice AI pipelines** and real-time audio processing

### Preferred Experience:
- Conversational AI and flow management systems
- OpenAI API integration
- Async Python programming
- Docker and dev containers
- Git version control

## Project Scope

### Phase 1 - Core Fixes (Immediate)
- Review and optimize existing `conversation_flow_graph.py`
- Fix integration between LangGraph and LiveKit voice pipeline
- Ensure agent speaks responses correctly
- Implement proper node transition logic

### Phase 2 - Enhancement (Following)
- Add advanced edge conditions and logic
- Implement context management between nodes
- Add support for multiple LLM providers
- Optimize performance and error handling

## Technical Stack
```
- LiveKit Agents
- LangGraph
- LangChain
- Pydantic
- OpenAI API
- Docker (dev container environment)
```

## Project Files Structure
```
conversation_flow_graph.py # Main implementation file
schemes.py # Pydantic schemas (DO NOT MODIFY)
external_services/ # Service integrations
logger.py # Logging setup
requirements.txt # Dependencies
```

## Deliverables
1. **Working conversation flow system** that processes nodes and speaks responses
2. **Optimized code** with proper LiveKit integration
3. **Documentation** explaining the solution and any architectural decisions
4. **Testing examples** to validate the implementation

## Timeline
- **Quick assessment**: 1-2 days to review code and provide initial feedback
- **Core implementation**: 5-7 days for main fixes
- **Testing and refinement**: 2-3 days

## Budget Range
**$500 - $1,500** depending on experience and scope completion

## How to Apply

### Include in Your Proposal:
1. **Experience summary** with LiveKit and LangGraph projects
2. **Approach overview** - how you would tackle this project
3. **Timeline estimate** for Phase 1 completion
4. **Questions or clarifications** about the requirements
5. **Portfolio examples** of similar voice AI or conversation flow work

### Sample Questions to Address:
- Have you worked with LiveKit Agents before?
- Experience with LangGraph state machines?
- How would you approach debugging voice pipeline integration issues?
- Any experience with conversational AI flow systems?

## Additional Information
- This is a **dev container environment** on Debian Linux
- All tools and dependencies are pre-configured
- You'll have full access to the codebase via Git
- We prefer **clean, well-documented code**
- **Communication is key** - we value developers who ask good questions

## Next Steps
After selection, we'll provide:
- Complete codebase access
- Detailed technical documentation
- Live session to walk through current implementation
- Clear success criteria and testing procedures

---

**Looking for someone who can start immediately and deliver high-quality results. If you have the right experience and can commit to the timeline, we'd love to hear from you!**

Project Description:
We are looking for an experienced developer or development team to build a scalable, white-label-ready RAG (Retrieval-Augmented Generation) chatbot system with the following key features:

Project Goals:
- Develop a multi-tenant chatbot platform (SaaS-style)
- Deploy the solution in a GDPR-compliant EU-based hosting environment
- Integrate a LLM model (e.g., OpenAI, Claude, Mistral) with RAG capabilities
- Build a custom web interface that allows each client to:
-- Log into their own dashboard
-- Upload their own documents (PDF, DOCX, etc.)
-- Use those documents as a private data source for the chatbot
-- Access and view their own chat history
-- See the original document source for each chatbot response (with citation or link to the file section)

Technical Requirements:
Backend: Node.js or Python (LangChain, LlamaIndex, or Haystack preferred)
Vector Database: Weaviate, Qdrant, or Chroma (self-hosted or EU-based)
LLM Integration: OpenAI via Azure EU, Claude via AWS Frankfurt, or local models (e.g., Mistral, Llama 3)
Hosting: Hetzner, IONOS, or any GDPR-compliant EU-based provider
Frontend: Next.js or React with login, file upload, chat component, and tenant-specific views
Authentication: JWT or OAuth
Optional: Admin panel for managing tenants

Nice to Have:
Proven experience with RAG and vector search systems
Knowledge of GDPR compliance and data protection
Experience in multi-tenant SaaS architecture
Clean, maintainable code with documentation and containerization (e.g., Docker)

Deliverables:
Fully functional prototype including:
Multi-tenant login system
Secure file upload and indexing per client
RAG-enabled chatbot interface with source citation for every answer
GDPR-compliant deployment setup
Technical documentation and deployment instructions

We are planning for long-term collaboration and ongoing development for additional client projects.
We look forward to your application!

We’re looking for a skilled full-stack developer to help us build a scalable, templated system for deploying tens or even hundreds of lightweight landing page websites in the home improvement niche (plumbing, HVAC, locksmith, etc.).

Each site will be simple (like a one-page landing site), but personalized by location and possibly brand. The system should allow easy duplication and customization of each website via config files or a CMS-like interface. You’ll build the initial boilerplate and help us scale from there.

⸻

Key Responsibilities:
• Build a reusable landing page boilerplate using Next.js or Astro
• Implement dynamic content loading via JSON, YAML, or Airtable (e.g., city, service type, contact details)
• Integrate IP-based geolocation to customize on-page headlines and content dynamically
• Set up clean, responsive styling (preferably using Tailwind CSS)
• Connect forms to external services (e.g., Zapier or webhook URLs)
• Set up deployment pipelines (e.g., Vercel, GitHub Actions) to allow easy rollout of new pages
• Support custom domain/subdomain configuration (e.g., using Cloudflare)

⸻

Ideal Experience:
• Expert in React/Next.js or Astro
• Strong understanding of static site generation and SEO
• Experience with IP-based personalization (ipinfo.io, ipapi.co, etc.)
• Familiar with CI/CD tools (GitHub Actions or similar)
• Solid grasp of responsive design (mobile-first, fast loading)
• Experience integrating 3rd-party form tools (Zapier, Formspree, Getform, etc.)
• Bonus: Experience working with Airtable or Notion APIs

⸻

Optional Bonus Skills:
• Automated DNS/domain setup via Cloudflare API
• Experience with headless CMS setups
• Knowledge of local service SEO best practices

⸻

Deliverables (Phase 1):
• Fully working boilerplate landing page
• Admin/config system for generating new sites (JSON/Airtable)
• Sample deployments for 2–3 example cities
• Geo-personalization logic implemented

⸻

Ongoing Work:
• Launching and managing 10–100+ landing pages
• Periodic improvements to the system (speed, UX, SEO)
• New feature support (live chat, analytics, etc.)

⸻

To Apply:
Please share:
• Relevant portfolio links (especially landing pages)
• What framework you’d use and why (Next.js, Astro, etc.)
• A quick summary of how you’d approach scaling this system
• Estimated time/cost for the initial boilerplate (not binding)

We are seeking a skilled web developer to create a simple, non-KYC crypto swap website. The platform should allow users to easily swap various cryptocurrencies without the need for identity verification. An intuitive admin panel is also required for managing transactions and user activities. The ideal candidate will have experience in blockchain technology, web development, and a strong understanding of cryptocurrency functionalities.

If someone even has a white label that I can just change the branding that would be perfect too.

Examples:
ChangeNow.io
Changehero.io
SimpleSwap
StealthEx

# AI-First Platform Developer: Email + AI Concierge MVP (Smart Life Assistant)

## Revolutionary Vision
We're building an **AI Concierge Platform** that manages your entire digital life. Think "Jarvis from Iron Man" meets modern email service. Our AI analyzes your emails and services to proactively manage your life, detect conflicts, and provide intelligent suggestions.

## Core Innovation:
**AI Concierge = Our Main USP**
- Multi-LLM system: Route tasks to best/cheapest model for each job
- Smart email analysis: Extract events, deadlines, important info automatically
- Proactive conflict detection: "Meeting conflicts with doctor appointment"
- Cross-service intelligence: Email insights inform other services

## MVP Strategy:
**Phase 1**: Email + Basic AI Assistant (validate AI concept)
**Phase 2**: Advanced AI Concierge (multi-LLM routing, smart suggestions)
**Phase 3**: Multi-service integration (VPN, calendar, etc.)

## Technical Requirements

### Core Stack:
- **Backend**: Python (FastAPI) - optimal for multi-LLM integration
- **AI/ML**: Multi-LLM routing system (LLM via Openrouter)
- **Database**: Vector database (Pinecone/Weaviate) + PostgreSQL
- **Email API**: Mailgun/Resend for reliable email delivery
- **Frontend**: React/Next.js with real-time AI chat interface
- **Mobile**: React Native for instant AI notifications

### AI Development Expertise (CRITICAL):
- **Quality-Cost Optimization** - Balancing AI performance with operational efficiency
- **Dynamic Model Selection** - Real-time routing based on quality requirements
- **AI Performance Monitoring** - Tracking quality metrics across different models
- **Email/Data Processing** - Complex email content parsing with high accuracy
- **Vector Databases** - Semantic search and content indexing
- **Quality-Aware Systems** - Building AI that prioritizes user satisfaction
- **Full-stack AI Applications** - Complete AI product development with quality focus

### Modern Development Tools:
- **AI-Assisted Coding**: GitHub Copilot, Claude, Cursor for rapid development
- **LLM API Management**: Experience with multiple model providers
- **Cost monitoring**: Track LLM usage and optimize spend

## Project Scope

### Phase 1 - Email + Basic AI:
**Email Foundation:**
- Clean email service (send/receive via Mailgun)
- Modern web interface with AI chat sidebar
- Mobile-responsive design

**Basic AI Concierge:**
- Multi-LLM routing system (LLM1 for analysis, LLM2 for summaries)
- Email summarization and priority detection
- Simple AI commands and natural language queries
- Context-aware responses

**Smart LLM Architecture:**
- Quality-first approach with cost optimization
- Vector database for efficient email search
- Background processing for email analysis

### Phase 2 - Smart AI Concierge:
**Advanced AI Features:**
- **Dynamic LLM Selection**: Real-time model selection for optimal quality-to-cost ratio
- **Context-aware routing** balancing performance and efficiency
- **Quality-first approach** with cost optimization
- **Smart escalation chains** when quality matters most

**Quality-First Model Selection:**
- Real-time quality-cost analysis for model selection
- Performance monitoring and quality score tracking
- Smart escalation when quality thresholds aren't met
- User satisfaction optimization with cost awareness

### Phase 3 - Multi-Service Integration:
**VPN Integration Preparation:**
- Database schema for VPN service integration
- User model extension for service permissions
- API endpoints planning for future VPN connection
- Architecture documentation for seamless service addition

## AI Concierge Use Cases

### Email Intelligence Examples:
```
📧 Bills & Finance:
"Your electricity bill is due in 3 days, but I notice you're traveling.
Set up auto-pay or pay now?"

📧 Health & Appointments:
"Lab results came in - everything normal. Your next checkup is in 2 months,
should I remind you to schedule?"

📧 Work & Deadlines:
"Project deadline moved to Friday, but you have 3 meetings that day.
Suggest rescheduling the 2PM call?"

📧 Social & Events:
"Wedding RSVP due tomorrow. Based on your calendar, you're free that weekend.
Should I confirm attendance?"

📧 Shopping & Subscriptions:
"Amazon package delayed, but you ordered it for tomorrow's party.
Want me to find same-day delivery alternatives?"

📧 Family & Personal:
"Mom's birthday is next week. Last year you sent flowers on Tuesday.
Similar gift this year or try something new?"
```

### Smart Automation Examples:
```
📊 Email Categorization:
- Bills → Auto-extract amounts, due dates
- Appointments → Add to calendar, set reminders
- Receipts → Track expenses, warranty info
- Social → Priority based on relationship strength

🔔 Proactive Notifications:
- "Credit card statement shows unusual spending in restaurants - reviewing budget?"
- "Three job applications sent last month, no responses yet. Time to follow up?"
- "Gym membership expires next month, but you haven't been in 3 weeks"

🎯 Context-Aware Suggestions:
- "Working late again? Your dinner reservation is in 2 hours"
- "Ordered coffee machine, but also signed up for coffee subscription yesterday"
- "Meeting with John tomorrow - here's context from your last 3 conversations"
```

### Dynamic Multi-LLM Routing:
```
Quality-First Model Selection → Optimal Cost-Performance Balance:

"Summarize my emails"
→ Llama 4 delivers 95% quality at $0.0001/1k tokens ✅

"I have a complex schedule conflict with multiple meetings and travel"
→ GPT-4 needed for 98% accuracy on complex reasoning ($0.03/1k tokens)
→ DeepSeek might only achieve 75% accuracy - quality gap too large ❌

"Quick question about my Amazon order"
→ DeepSeek delivers 90% quality at $0.0002/1k tokens ✅

"Help me write a sensitive email to my boss about promotion"
→ GPT-4 required for nuanced communication (quality critical) ✅
→ Cheaper models risk career-damaging mistakes ❌

Quality-Cost Decision Matrix:
- High-stakes situations: Quality trumps cost (use best model)
- Routine tasks: Cost-effective models that maintain acceptable quality
- User satisfaction: Monitor response quality scores per model
- Smart fallback: Auto-escalate if cheaper model fails quality threshold
- A/B testing: Continuously optimize model selection algorithms
```

## Ideal Developer/Agency Profile:

### Must Have Experience:
- **Quality-Cost Optimization** - Balancing AI performance with operational efficiency
- **Dynamic Model Selection** - Real-time routing based on quality requirements
- **AI Performance Monitoring** - Tracking quality metrics across different models
- **Email/Data Processing** - Complex email content parsing with high accuracy
- **Vector Databases** - Semantic search and content indexing
- **Quality-Aware Systems** - Building AI that prioritizes user satisfaction
- **Full-stack AI Applications** - Complete AI product development with quality focus

### Bonus Expertise:
- Personal assistant/productivity app development
- Natural language processing for personal data
- Cost optimization for AI/ML systems
- Background job processing systems
- Mobile AI app development

## Application Requirements:

### Must Include:
1. **Multi-LLM Portfolio**:
- Experience with multiple AI model providers
- Cost optimization strategies you've implemented
- Performance comparisons between different LLMs

2. **Technical Architecture**:
- How would you design the multi-LLM routing system?
- Cost-effective AI processing architecture
- Real-time AI response system design
- Background email analysis workflow

3. **AI Development Approach**:
- Task-specific LLM selection strategy
- Vector database setup and optimization
- Context management for conversational AI
- Cost monitoring and optimization methods

4. **Practical Examples**:
- Show examples of AI task routing you've built
- Demonstrate understanding of different LLM strengths
- Cost analysis of different AI approaches

## Key AI Features Checklist:

### Email Intelligence:
- [ ] Multi-LLM email analysis system
- [ ] Automatic categorization and priority detection
- [ ] Contact relationship mapping
- [ ] Content extraction (dates, amounts, deadlines)

### Intelligent Assistance:
- [ ] Conflict detection across different life areas
- [ ] Proactive suggestion generation
- [ ] Context-aware conversation memory
- [ ] Smart notification timing

### Quality-Optimized Dynamic AI:
- [ ] Real-time quality-cost analysis for model selection
- [ ] Performance monitoring and quality score tracking
- [ ] Smart escalation when quality thresholds aren't met
- [ ] User satisfaction optimization with cost awareness

### User Experience:
- [ ] Natural language email queries
- [ ] Real-time AI responses
- [ ] Learning user preferences
- [ ] Cross-service intelligence

## Critical Questions:
1. **Quality-cost optimization** - How do you balance AI quality with operational costs?
2. **Model performance tracking** - Show examples of quality monitoring systems you've built
3. **Smart escalation logic** - When and how do you automatically upgrade to better models?
4. **Quality threshold management** - How do you ensure minimum quality standards per task type?
5. **A/B testing for AI** - Experience with optimizing model selection based on user satisfaction?
6. **Practical examples** - Demonstrate quality-cost decision making in previous AI projects

## Success Metrics:
- **Optimal quality-cost balance** - Maximum user satisfaction at sustainable costs
- **Quality-first user experience** - AI that never compromises on critical tasks
- **Smart cost optimization** - Efficient spending without sacrificing quality
- **Performance monitoring** - Continuous quality improvement across all models
- **Scalable quality standards** - Consistent excellence as user base grows

**This is not just another email service - we're building the future of AI-powered personal assistance with smart cost optimization.**

---

*Only apply if you have real multi-LLM integration experience. Show us your AI portfolio and cost optimization strategies.*

We are seeking an experienced chatbot developer proficient in n8n to create and integrate automated workflows with AI-driven chat solutions. The ideal candidate will have a strong understanding of chatbot frameworks and be able to design user-friendly interactions. You'll be responsible for building, testing, and deploying chatbots that enhance user engagement and streamline communication. If you are passionate about automation and have a proven track record in chatbot development, we'd love to hear from you!

The ideal candidate will have strong experience with Office.js, Microsoft 365 integrations, and cloud backend development using Azure Functions. Experience with Microsoft Graph and AI/NLP services is highly desirable. Background working with education or public sector clients is a plus.

🔧 Key Responsibilities:
Develop a Microsoft Outlook add-in compatible with both desktop (Windows) and web clients

Build a clean, intuitive front-end UI using React, Vue, or a similar framework with Office.js

Design and build scalable, serverless back-end services using Azure Functions

Integrate with Microsoft Graph API, Azure Active Directory (AAD) authentication, and cloud storage like SharePoint or OneDrive

Develop customizable dashboards for administrators to monitor activity, view tagging history, and manage configurations

Incorporate AI-powered natural language processing (NLP) services for content scanning and classification

Ensure secure authentication, user preferences management, and compliance with data privacy standards

✅ Must-Have Skills:
Proven experience with Office.js or VSTO add-in development

Strong front-end skills with JavaScript, TypeScript, React, Vue, or similar

Back-end development experience with Azure Functions and serverless architecture

Microsoft Graph API integration expertise

Experience with secure authentication protocols (OAuth 2.0, MSAL)

Experience developing Outlook add-ins for both desktop and web platforms

🌟 Nice-to-Have:
Familiarity with K–12 compliance requirements

Experience with Power Automate, SharePoint Online, or Microsoft Teams integrations

Background in building SaaS tools or MVP products

Strong communication skills to work effectively with non-technical stakeholders

We're looking for a technical partner (agency or highly capable team) to help bring an AI + BI insights platform to life. This will be a customer-facing dashboard that aggregates and interprets data from 20+ public sources (e.g., retailer sites, industry reports, social media, scientific databases), which is done through scraping or AI ingesting particular sites and delivers actionable insights for consumer brands.

The platform will:

*Scrape and update data regularly (e.g., Sephora, Ulta, brand websites, review sites, social media)
*Interpret consumer sentiment, product trends, gaps in brand portfolios, market share insights
*Display insights in modular dashboards
*Enable comparisons across brands, categories, and time (e.g., week-over-week changes)
*Use generative AI to summarize reviews, trends, and recommendations

You'll be working with a product team that will provide detailed mockups and insight logic. We’re looking for a group that brings strong data warehousing, web scraping, AI integration, and front-end development capabilities.

📽️ A full video walkthrough is included in the post—please review it before responding. - https://www.loom.com/share/c8e4f1d0e3fe40cba51e5688f0fc52ea?sid=67cdd564-10f6-4bbf-9954-f90ed6c839dc

Looking for:
✅ Agency/Team with full-stack capabilities - MUST HAVE comparable experience.
✅ Experience building data-driven platforms (across multiple data sources)
✅ Ability to move fast and iterate
✅ Will look at independent individuals if you have teams or other partners you can bring with you.

MVP:
* Filter at the top to filter brands, products, and/or category and then hit "analyze" to create AI insights for selected criteria
* data warehouse well architected and organized
* ~20+ sites scrapped, and more added to the pool over time
* 5-10 strong insights surfaced per page for 2-3 pages

Note this is MVP which I want to launch as fast as possible and then will be continuing to grow the data library and pushing out more insights as we go.

**DO NOT EMAIL ME OR CALL ME OUTSIDE OF UPWORK - I WILL NOT WORK WITH THOSE THAT DO**

Please include in your response:

* Relevant examples of similar projects
* What tool you plan to use that could allow us to leverage open source code and move quick (e.g. do you have something for data visualization you prefer using for good looking visuals?)
* How you'd approach an MVP

We are looking for two types of people:

1. Those that can work Saturday and Sunday regularly
2. Those who can work fulltime Monday - Friday (8 hours/day)

Tech stack:

* Kotlin + Spring boot
* Thymeleaf for server side rendering
* Web awesome for UI web components: webawesome.com
* Web awesome is using Lit under the hood, which we also use for custom components not provided by the framework
* Postgres as DB
* Flyway for migrations
* Kamal for deployment to Ubuntu VPS (https://kamal-deploy.org)

You should be able to start the app locally in Intellij, but most importantly, you should master CSS, HTML, javascript at an advanced levels and have good taste and design skills.

Good backend skills for Java/Kotlin is not needed, but a bonus if you also master these.

Take a look at:

* https://webawesome.com
* https://lit.dev

Would you be able to:

1. customize the framework components so they look good for our app
2. create custom components that adheres to design and solve custom needs
3. Use modern and efficient CSS to make code maintainable