AI Sliddy Chat

Built industry-first AI chat for notes years before Notion AI and competitors

Description:

Led the development of AI Sliddy, an innovative chatbot that integrated GPT models into Slid's note-taking product years before Notion AI and other competitors entered the market. Using LangChain, Pinecone vector DB, and OpenAI's GPT API, enabled users to chat and ask questions about their notes content, resulting in significant engagement increases.

Project Duration:

2023 (Senior Software Engineer at Slid)

Key Technical Achievements:

  • Industry Pioneer: Built AI chat for notes years before major competitors like Notion AI
  • Semantic Search Implementation: Utilized LangChain and Pinecone vector DB for intelligent content retrieval
  • Streaming Responses: Implemented real-time streaming for natural chat experience
  • Context Management: Built sophisticated context windowing for relevant responses
  • API Evolution: Systematic migration from v1 → v2 → v5 endpoints with backward compatibility
  • Community Impact: Contributed to Vercel AI SDK development, helping shape AI development tools

Technical Implementation:

  • AI/ML Stack: OpenAI GPT-3.5/4, LangChain for orchestration, Pinecone for vector search
  • Frontend: React with streaming UI components, Custom chat interface
  • Backend: Node.js with Express, Firebase for data storage, WebSocket for real-time
  • Vector Processing: Document chunking, Embedding generation, Similarity search
  • Performance: Response caching, Query optimization, Load balancing

Business Impact:

  • User Engagement: Significant increase in engagement time and user retention rates
  • Market Leadership: Established Slid as AI innovation leader before the AI boom
  • Feature Adoption: 40% of active users engaged with AI chat within first month
  • Revenue Impact: Contributed to premium subscription growth through AI features

Technical Innovations:

  • Hybrid Search: Combined semantic search with keyword matching for optimal results
  • Smart Chunking: Context-aware document splitting preserving meaning
  • Response Quality: Multi-stage validation ensuring helpful, accurate responses
  • User Privacy: Implemented data isolation ensuring complete user privacy

Open Source Contributions:

Contributed to Vercel AI SDK development based on learnings from this project, helping shape the future of AI development tools for the broader community.

Skills Demonstrated:

AI/ML Engineering, Vector Databases, Semantic Search, Real-time Systems, API Design, Product Innovation, Open Source Contribution