Hill Patel (
STiFLeR7) — AI Developer & Researcher at CloudRedux, based in Pune, India 🇮🇳. I build open-source Agentic AI systems and specialize in Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), the Model Context Protocol (MCP), Edge AI, and model quantization — shipping scalable intelligence that runs on real hardware, from cloud GPUs down to 6GB consumer cards and on-device inference.🔭 Currently building
memex— persistent, bitemporal memory for AI coding agents over MCP (Tree-sitter + Gemini → Neo4j).
classHill_Patel(AI_Researcher):
""" [INFO] Building open-source AI tools for the community. [WARN] High compute requirements detected. """def__init__(self):
self.code="STiFLeR7"self.specs= {
"role": "AI Developer & Researcher @ CloudRedux",
"focus": ["Agentic AI", "MCP", "RAG", "Edge AI", "Quantization"],
"driver": "Deploying Scalable Intelligence"
}
defexecute_mission(self):
whileTrue:
self.research()
self.optimize()
self.deploy("Production")
🧠 Agentic AI & Retrieval · MCP · RAG · LangGraph · Vector Search · Neo4j · Knowledge Graphs
⚡ Model Optimization · Quantization · ONNX · TensorRT · CUDA · Pruning
-ff6b6b?style=flat-square&logo=compress&logoColor=white){kind=icon}
🔬 ML Engineering · Transformers · CNNs · Multimodal · PyTorch · TensorFlow
🚀 Backend & MLOps · FastAPI · Docker · GitHub Actions · Redis · GCP · AWS
| PROJECT | MISSION BRIEF | CORE TECH |
|---|---|---|
| 🧠 memex ⭐ | [MCP-MEMORY] Persistent, bitemporal knowledge graph of your codebase, served to Claude Code, Cursor & any MCP client. 14 MCP tools, Tree-sitter + Gemini → Neo4j. | MCPNeo4jGraphiti |
| 📦 imgshape ⭐ | [CV-TOOLKIT] Dataset intelligence toolkit — deterministic fingerprints, explainable decisions & production-ready artifacts for reproducible ML. | PythonPyPICV |
| 🔎 agentic-rag | [RAG-SYSTEM] Production-grade Agentic RAG that runs reliably on 6GB consumer GPUs (RTX 3050). Controllable, graph-based, explicit failure handling. | LangGraphRAGAgents |
| ⚡ DevPulseAIv3 | [PLATFORM] Cloud-native intelligence platform — multi-agent LLM pipeline aggregates dev signals into real-time chat, REST API & scheduled digests. | Multi-AgentFastAPILLM |
| 🛠️ prompt-forge ⭐ | [DEV-TOOL] Prompt refinement engine — packaged as a Claude Code plugin. | JavaScriptLLMMCP |
| 🤖 dex | [AI-ASSISTANT] Persistent, memory-aware personal AI assistant built on the Model Context Protocol. | MCPTypeScriptLLM |
- Published Research — Transforming Urban Solutions for Smart Cities through Crowdsourced Feedback (Mar 2025)
- MCP Mastery — Model Context Protocol · Fractal Analytics
- Professional Certificate — RAG and Agentic AI · Coursera
- Course Completion — Neural Networks with PyTorch · Coursera
- Building a Memory Layer for Claude Code, Cursor, and Gemini CLI
- Most Vision AI Systems Stop at Detection — I Built One That Takes Action
- Most RAG Systems Fail Quietly — Here’s How I Built a 98%-Accurate Agent on a 6GB GPU
- MedMNIST-EdgeAI: Compressing Medical Imaging Models for Efficient Edge Deployment
- LCM vs. LLM + RAG
