AI Engineer | Building Production-Ready LLM Systems & Intelligent Automations

I'm a Computer Science undergraduate specializing in AI & Machine Learning. I build production-grade LLM-powered systems, RAG pipelines, and AI-driven automations that deliver measurable impact.

• 🔬 Focus Areas: LLM Applications, RAG Systems, Agentic Workflows, AI Automation
• 🏆 1st Place Winner - AI Demos × AI For Content Creators Hackathon
• 🛠️ Building: Evaluation-driven GenAI solutions with LangChain, n8n, and custom agent pipelines

Conversational Q&A over YouTube videos with 40-50% hallucination reduction

• Built with LangChain, FAISS, Hugging Face LLMs, and FastAPI
• Implements LLMOps with RAGAS evaluation (context relevance, faithfulness, answer relevancy)
• Features semantic chunking, multi-turn reasoning, and Chrome extension integration

AI content automation reducing creator workload by ~80%

• Repurposes long-form videos into publish-ready formats using Gemini video analysis
• Features fact-grounding, engagement scoring, and multi-platform optimization
• Built with Streamlit, Google Gemini, FFmpeg, and SQLite

AI-powered patient-doctor communication bridge with dual-view generation

• Generates patient-friendly summaries (Hindi/English) and structured SOAP notes
• Local AI privacy with Llama 3, role-based access control, PostgreSQL backend
• Built with Flask and Ollama for on-premise deployment

Real-time drowsiness detection with ResNet-18 and live webcam inference

• Edge and IoT-ready system for driving and industrial safety
• Computer vision pipeline with face detection and temporal tracking
• Built with PyTorch, ResNet-18, and OpenCV

Automated UX evaluation combining HTML analysis, Google Analytics, and AI-powered heuristics

• Evaluates websites against 8 core UX principles (clarity, accessibility, cognitive load)
• Generates severity-based findings with prioritized recommendations
• Built with n8n, Claude AI, Google Analytics API

Transforms client transcripts into production-ready video storyboards

• Generates scene-by-scene breakdowns with Figma-ready UI specs
• Includes quality assurance scoring and realistic sample data
• Built with n8n, OpenAI GPT-4, and JavaScript

End-to-end AI blog automation with research, writing, and publishing

• Deep web research via Tavily, SEO optimization, and Sanity CMS integration
• Generates long-form posts with professional hero images (Gemini)
• Built with n8n, OpenAI, Tavily, and Sanity CMS

I'm always interested in collaborating on AI projects, discussing LLM applications, or exploring new opportunities in AI engineering.

• 🌐 Portfolio: garvit-portfolio.vercel.app
• 💼 LinkedIn: linkedin.com/in/garvit-audichya
• 📧 Email: audichyagarvit010@gmail.com
• 💻 GitHub: You're already here! Check out my repositories