I am a Machine Learning Engineer focused on building production-grade AI systems. My expertise spans Agentic Workflows, RAG Pipelines, and scalable MLOps Infrastructure. I bridge the gap between research code and deployed microservices.
🔭 Currently exploring: Agentic Agents, Model Context Protocol (MCP), and Multi-Agent Orchestration.
👯 Looking to collaborate on: End-to-End GenAI platforms and Open Source MLOps tools.
🌱 Learning: Advanced Model Quantization (Unsloth) and Rust for ML.
👯 Looking to collaborate on: End-to-End GenAI platforms and Open Source MLOps tools.
🌱 Learning: Advanced Model Quantization (Unsloth) and Rust for ML.
| Generative AI & LLMs |
|
| MLOps & Cloud |
|
| Core AI/ML |
|
| Backend & Viz |
|
|
Cost-Optimized LLM Routing System Intelligent router using ML to direct queries to optimal models (Llama 8B vs 70B), reducing costs while maintaining quality. 🛠️ Python, LangChain, Random Forest, ChromaDB, FastAPI, LLM APIs, Streamlit |
|
|
End-to-End MLOps Forecasting Production forecasting pipeline with real-time drift detection, feature stores, and Kubernetes deployment. 🛠️ Python, XGBoost, LightGBM, Optuna, scikit-learn, DVC, Evidently, Docker, Kubernetes, Streamlit, GitHub Actions |
|
|
Real-Time Fraud Detection API High-throughput banking security system handling class imbalance, fully tracked via MLflow and deployed on AWS. 🛠️ Python, XGBoost, LightGBM, Optuna, MLflow, DVC, Docker, FastAPI, AWS, GitHub Actions |
|
|
Deep Learning Medical Imaging CNN-based tumor classification pipeline utilizing ResNet architectures and OpenCV for automated diagnosis support. 🛠️ Python, TensorFlow, Keras, ResNet, OpenCV, MLflow, DVC, Docker, Streamlit, scikit-learn |
