Welcome to my GitHub! I'm a Senior AI/ML Engineer with a decade of experience specializing in Agentic AI, LLM Engineering, and Enterprise AI Solutions at scale.

• LLM Engineering: Custom model fine-tuning (LoRA, QLoRA), RAG architectures, prompt engineering
• Agentic AI: Multi-agent systems with LangChain, autonomous workflows, tool-use implementations
• Production AI: MLOps pipelines, model versioning, A/B testing, monitoring at scale
• Vector Systems: FAISS, Pinecone, ChromaDB for semantic search and retrieval

• Multi-Agent Architecture: Building agent-based systems for context management and iterative code generation
• RAG Optimization: Implementing production-ready retrieval systems with 75% faster document retrieval
• LLM Fine-tuning: Achieved 35% performance improvement on domain-specific tasks with 40% model size reduction

frameworks = [
    "PyTorch", "TensorFlow", "Transformers", 
    "Lang - Graph & Chain", "LlamaIndex", "PEFT", "scikit-learn"
]

models = [
    "GPT-4", "Claude-4", "Llama-3", "DeepSeek", 
    "Mistral", "Phi-3", "Gemini", "CodeLlama"
]  

vectorDBs = [
    "FAISS", "Pinecone", "Weaviate", 
    "ChromaDB",
]

tools = [
    "OpenAI", "Anthropic", "Azure OpenAI",
    "Ollama", "HuggingFace", "MLflow", "WandB"
]

• Azure: AI Services, Cognitive Services, Function Apps, AKS, Cosmos DB
• AWS: SageMaker, EC2, S3, Lambda, DynamoDB
• Ops: Docker, Kubernetes, MLflow, Weights & Biases

• Microsoft Azure Fundamentals (AZ-900)
• Azure AI Fundamentals (AI-900)
• Azure Data Fundamentals (DP-900)
• Power Platform Fundamentals (PL-900)
• Executive PG Program in AI & ML

• Backend: Python (FastAPI, Flask), C# (.NET Core), Node.js
• Frontend: React, Angular, TypeScript, Tailwind CSS, Three.js
• Databases: SQL Server, PostgreSQL, MongoDB, DuckDB, Redis

Production-ready multi-agent RAG system with autonomous document processing and context management.
Tech: LangChain · FAISS · GPT-4 · FastAPI

End-to-end pipeline for fine-tuning LLMs with LoRA/QLoRA on custom datasets.
Tech: PyTorch · Transformers · PEFT · MLflow

Custom language model training on TinyStories dataset with attention mechanisms.
Tech: PyTorch · HuggingFace · WandB

Metadata-driven document retrieval using LLM extraction instead of traditional chunking.
Tech: LangChain · OpenAI · Vector DBs · Streamlit

Educational platform with structured deep learning curriculum and interactive examples.
Tech: JavaScript · HTML/CSS · D3.js

Production chatbot with RAG, memory management, and tool integration.
Tech: Python · LangChain · FastAPI · Redis

Automated AI workflows for document generation, code review, and DevOps automation.
Tech: n8n · Claude/GPT-4 · GitHub Actions

Novel ML classifier using magnetic force principles for decision boundary optimization.
Tech: scikit-learn · NumPy · Jupyter

AI-powered tool to analyze and visualize large codebases with dependency mapping.
Tech: Python · AST · NetworkX · Plotly

CPU: AMD Ryzen 9 9950X (16-core, 32 threads)
GPU: NVIDIA GeForce RTX 5090 (32GB VRAM)
RAM: 96GB DDR5 @ 5600MHz
Storage: Samsung 990 PRO 2TB NVMe
Purpose: LLM training, fine-tuning, inference

CPU: AMD Ryzen 9 7900X (12-core, 24 threads)
GPU: NVIDIA RTX 4070 SUPER (12GB VRAM)
RAM: 32GB DDR5
Purpose: Distributed training, model serving

• 🤖 Agentic AI Systems - Building autonomous agents for enterprise applications
• 🧬 LLM Optimization - Fine-tuning and deploying efficient models at scale
• 📊 RAG Architecture - Advanced retrieval systems with <100ms latency
• 🔄 MLOps Automation - End-to-end pipelines from training to production

• 🌐 Portfolio
• 💼 LinkedIn
• 📧 karthickrajam18@gmail.com
• 🏢 Current: Lead AI/ML Engineer @ Appian
• 📍 Location: Chennai, India

Interested in:

• 🤝 Open-source AI/ML projects
• 🧬 LLM fine-tuning and optimization
• 🏗️ Enterprise AI architecture design
• 📚 Knowledge sharing and mentoring

⚡️ "Building intelligent systems that augment human capability, not replace it"

Agentic AI LLM Engineering RAG Systems Multi-Agent Architecture PyTorch LangChain Transformers Vector Databases MLOps Fine-tuning Claude GPT-4 Llama FAISS Pinecone Azure AI AWS Bedrock Enterprise AI Production ML Full-Stack Cloud Architecture