About

About Me

I’m Jyoti Ranjan, a GenAI architect specializing in enterprise-grade AI systems, LLM fine-tuning, vLLM inference optimization, and Document AI. I focus on building production systems that are fast, reliable, and scalable.

Core Expertise

• LLM fine-tuning (SFT, RLHF, DPO)
• vLLM inference engineering
• OCR + Document AI (Donut, TrOCR)
• Multi-agent systems (Autogen, CrewAI)
• RAG + GraphRAG architectures
• Azure OpenAI, AWS Bedrock/SageMaker

Tech Stack

AI/ML Frameworks: PyTorch • TensorFlow • Deep Learning • NLP • Computer Vision

GenAI & LLMs: LangChain • RAG • GraphRAG • LLM Fine-tuning • RLHF/DPO • vLLM • Azure OpenAI • AWS Bedrock

Document AI: Donut • TrOCR • OCR Pipelines

Multi-Agent: Autogen • CrewAI • Agentic Frameworks

Big Data: Hadoop • Spark • Kafka • HBase • ELK Stack

Databases: Neo4j • Vector DBs

Cloud & DevOps: Azure • AWS • Docker • Kubernetes • MLFlow

Programming: Python • C/C++ • Linux System Programming