I build end-to-end AI systems that solve real-world problems from LLM-powered applications to sequence-based recommendation systems.
Focused on turning models into usable, scalable products, not just experiments.
- Design and build RAG-based AI systems with grounded outputs
- Develop Transformer-based recommendation models for sequential prediction
- Work on Computer Vision systems (YOLO, OpenCV) for real-world automation
- Build backend APIs using FastAPI for deploying AI models
- Built an end-to-end RAG pipeline using LLaMA + ChromaDB
- Enables natural-language queries over thousands of real reviews
- Produces evidence-backed insights (no hallucination)
👉 Focus: LLM reliability, semantic retrieval, real-world usability
- Designed a sequence-aware recommender using Transformer architecture
- Modeled user behavior like GPT-style next-token prediction
- Captures temporal patterns in user preferences
👉 Focus: Deep learning fundamentals, sequence modeling
- Built an intelligent chatbot using LangChain + FAISS + Gemini
- Retrieves relevant context before generating responses
- Supports dynamic knowledge base updates
- Developed a system to detect and extract structured fields from PDFs
- Applied computer vision techniques for real-world document processing
- Languages: Python, Java
- AI/ML: PyTorch, Transformers, YOLO, OpenCV
- LLM Stack: LangChain, Ollama, FAISS, ChromaDB
- Backend: FastAPI, REST APIs
- Tools: Git, Streamlit
AI/ML Intern — Param Group of Companies
- Built production-oriented AI systems across CV and LLM domains
- Worked on automation and real-world deployment pipelines
- I build complete systems, not isolated models
- Strong focus on practical AI (RAG, pipelines, APIs)
- Ability to connect ML concepts → real applications
- GitHub: https://github.com/Viidhii19