Aryan Shirke
AI/ML + LLM Engineer, API Designer, ML Researcher, FastAPI Developer, System Architect, GenAI Developer, Database Designer, LangChain Developer, and Model Optimizer
Cost-Efficient Deployment of a Graph RAG System on GCP with Neo4j and LangChain
A complete guide and estimation for deploying a Graph-based Retrieval-Augmented Generation (RAG) system using Neo4j, LangChain, and Azure OpenAI on Google Cloud Platform (GCP). This setup is optimized for scale (10,000+ nodes), and includes architecture planning, schema ingestion, query generation using LLMs, database sizing, and detailed GCP cost breakdown for minimum and scalable deployments.
AI-Powered Document Extraction System
Engineered an AI-based pipeline leveraging OCR (PaddleOCR), YOLOv8, and ResNet50 to accurately extract tabular and metadata information from industrial PDF reports for companies like JSW, Tata, and Hyundai Steel.
Designing intelligent surveillance with YOLOv11
Building a high-accuracy person detection system for video surveillance using YOLOv11, GPU acceleration, and custom training pipelines.
AI-Powered Document Extraction System
Engineered an AI-based pipeline leveraging OCR (PaddleOCR), YOLOv8, and ResNet50 to accurately extract tabular and metadata information from industrial PDF reports for companies like JSW, Tata, and Hyundai Steel.
Hi there
I’m Aryan Shirke, an AI/ML + Backend Engineer currently working at Mahindra & Mahindra. I specialize in building intelligent systems, scalable APIs, and end-to-end GenAI applications. My work blends machine learning, backend infrastructure, and rapid prototyping to deliver impactful solutions.
In my free time, I enjoy exploring state-of-the-art models, building research-backed tools, and mentoring juniors. I’m always open to exciting opportunities or collaborations—feel free to get in touch.
Send me a message