Tech Stack: Node.js, Next.js, TypeScript, Zod, Atlas MongoDB, Firebase Auth
• Created a chatbot using RAG(Retrieval-Augmented Generation), OpenAI GPT-4o AI, vector embedding, Atlas MongoDB, and WebSocket, reducing resolution time for common user queries by 30%.
• Developed a referral system using web browser cookies to store the token, increasing the client base by 20%.
• Improved Information Retrieval features with Gemini 1.5 pro AI by using concurrent processing and sending all the Api requests all at once, reducing the latency by 2x.
• Wrote technical documentation for the code base and made the onboarding process 2x faster.
• Enhanced document filtration feature using Gemini 1.5 pro AI that filtered out unnecessary pages from the user’s legal document, reducing document review time for attorneys by 30%.
• Optimized LLM’s prompts to reduce the total cost for each API request.
• Engineered a multi-document scanning and consolidation feature, reducing average user task completion time for uploading by 60%.