United States
AI/ML Engineer with 4+ years of experience designing and deploying scalable AI, Machine Learning, and Data-Driven solutions. Specialized in Generative AI, Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Agentic AI, and MLOps. Proven expertise in building end-to-end ML pipelines, production-grade AI systems, and cloud-native solutions using AWS, Python, LangChain, LangGraph, and modern AI frameworks. Experienced in developing intelligent applications leveraging LLMs, Vector Databases, Knowledge Graphs, GraphRAG, and Multi-Agent Architectures to solve complex business challenges. Skilled in model fine-tuning, prompt engineering, scalable deployment, and AI governance. Passionate about transforming innovative AI research into impactful real-world solutions that drive automation, efficiency, and business growth. Core Skills: Generative AI | LLMs | RAG | GraphRAG | Agentic AI | LangChain | LangGraph | Python | AWS | MLOps | Vector Databases | Knowledge Graphs | Machine Learning | Deep Learning
• Designed and deployed an AI-powered payment intelligence platform using GraphRAG, Knowledge Graphs, GNNs, and LLMs for fraud detection, transaction risk scoring, and merchant monitoring, improving fraud detection precision by 31%. • Built real-time data pipelines on AWS SageMaker, Kinesis, Glue, Spark, and Redshift processing 120M+ payment records to support transaction intelligence and risk analytics. • Developed Agentic AI solutions using LangGraph, CrewAI, AutoGen, and MCP, reducing fraud investigation time by 37% through autonomous multi-agent workflows. • Implemented GraphRAG architectures with Neo4j, vector databases, and hybrid retrieval systems to deliver explainable risk insights and compliance recommendations, improving investigator productivity by 29%. • Fine-tuned Llama, Mistral, and Transformer models using QLoRA, PEFT, RLHF, and DPO on large-scale payment datasets, achieving F1 Score of 0.87 and ROC-AUC of 0.91. • Established enterprise LLMOps and MLOps pipelines with MLflow, Kubeflow, Docker, Kubernetes, and Terraform, improving model reliability and deployment efficiency. • Implemented AI governance, model monitoring, drift detection, and compliance frameworks aligned with PCI DSS and SR 11-7 standards using AWS CloudWatch and SageMaker Model Monitor.
• Architected semantic retrieval platform using embedding models and vector databases for large-scale operational datasets, improving search relevance and reducing average query latency by 14%, enabling faster information discovery across distributed applications and workflows. • Orchestrated retrieval-augmented response pipelines processing 2M records across enterprise data sources, improving information accessibility and reducing manual troubleshooting effort by 29%, while delivering contextual insights supporting operations, customer experiences, and decisions. • Enhanced similarity scoring and ranking mechanisms through embedding optimization, relevance tuning, and evaluation frameworks, improving top-k retrieval precision by 17%, increasing response quality and retrieval effectiveness across search and recommendation ecosystems. • Implemented scalable data ingestion and transformation pipelines for structured and unstructured datasets, standardizing feature extraction workflows and improving indexing throughput by 21%, supporting high-volume data processing requirements across machine learning platforms. • Operationalized production machine learning inference services using Docker, Kubernetes, and CI/CD pipelines, streamlining release automation and reducing deployment cycle time by 37%, while ensuring scalability, reliability, and consistency across multiple environments. • Refactored legacy data processing and ML modules into microservice-based architectures, developing Python-driven evaluation frameworks that improved maintainability and reduced manual validation overhead by 26% across retrieval, ranking, and model assessment workflows. • Collaborated with cross-functional engineering and product teams to deploy scalable solutions, establishing monitoring, performance benchmarking, and experimentation practices that improved operational efficiency, accelerated feature delivery & boosted platform adoption.