I build end-to-end intelligent systems — from multi-agent swarms and RAG pipelines to on-device AI apps. Engineering student (2027 batch), focused on Agentic AI and production-grade LLM systems.
Open for remote roles and freelance projects in AI automation, LLM engineering, and GenAI.
- Agentic AI & RAG: LangGraph, LangChain, Google ADK, Multi-Agent Swarms, MCP (Model Context Protocol)
- AI/ML: PyTorch, YOLOv8, DenseNet, LLM Fine-Tuning (LoRA), Sentence Transformers, Pinecone
- Mobile: Flutter (Android/iOS), Dart, Firebase, Gemini on-device
- Backend: FastAPI, Node.js, Flask, PostgreSQL, SQLite
| Project | Description | Stack |
|---|---|---|
| AI Package Conflict Resolver | Multi-agent swarm that autonomously resolves Python dependency conflicts. Live MCP server on HuggingFace. | LangGraph, Gemini, Pinecone |
| Calendar AI | 100% on-device agentic scheduler. No backend, no cloud costs. | Flutter, Gemini 2.5 Flash |
| ArchMind | Generates professional cloud architecture diagrams from text or GitHub URLs. MCP server available via npx. | React, Gemini, Mermaid.js |
| Sleep Protection | Privacy-first Android app that silences calls during sleep with emergency bypass via SMS. Zero cloud. | Flutter, Kotlin |
| Currency Detector | Real-time Indian currency detection for visually impaired users. 93.5% YOLO accuracy + DenseNet CNN. | YOLOv8, TensorFlow, Flutter |
| Career Assistant | RAG-based AI mentor that analyzes resumes and generates personalized career roadmaps. | LangChain, Gemini, Pinecone |
- LeetCode: 191 solved (92 Easy / 96 Medium / 3 Hard) · Contest Rating 1528 · Top 36%
- Kaggle: yashwantk23cse · Indian Currency Dataset published
- Strong in: Dynamic Programming, Backtracking, Arrays
