Hire Aarthi J. - MLOps Engineer

* Leading Generative AI initiatives at Citi Group as a Gen AI ML Engineer, driving innovation in NLP, computer vision, and automation. I design, fine-tune, and deploy Large Language Models (LLMs) including GPT, T5, LLaMA, and Mistral for enterprise copilots and domain-specific solutions. * Built RAG (Retrieval-Augmented Generation) pipelines using FAISS & LangChain to enhance enterprise knowledge search. * Implemented LoRA/QLoRA fine-tuning, RLHF, and prompt optimization for domain adaptation. * Deployed diffusion & GAN models (Stable Diffusion) for text-to-image and document intelligence. * Scaled production AI systems on Azure with Docker, Kubernetes, FastAPI & Streamlit. * Applied quantization, pruning, and distillation for inference optimization and cost efficiency. * Integrated LLMs with external APIs and automation frameworks (LangChain, AutoGPT, CrewAI). * Established AI evaluation & monitoring frameworks, ensuring fairness, explainability (LIME, SHAP), and compliance. * My work bridges research with production, enabling Citi to deploy scalable, ethical, and high-impact AI systems that improve automation, decision-making, and customer experiences.

At Fidelity, I delivered full-stack AI solutions — from research and fine-tuning to deployment — for NLP, multimodal AI, and automation at scale. Built end-to-end AI systems for production rollout. Fine-tuned transformers (GPT, BERT, T5) for NLP/text generation. Developed retrieval-augmented AI (Pinecone, FAISS, Weaviate). Created AI copilots & assistants using LangChain, OpenAI, LlamaIndex. Designed & trained CNNs, RNNs, diffusion models for multimodal AI. Deployed AI on AWS with Docker, Kubernetes, FastAPI. Automated ML workflows with MLflow, Airflow, Ray Tune, W&B. Accelerated inference via quantization, ONNX, TensorRT, distillation. Engineered semantic search engines & embedding retrieval systems. Applied advanced NLP: summarization, sentiment, translation, classification. Built GPU-optimized inference APIs for low-latency solutions. Delivered AI-driven automation & predictive insights across domains. Applied LoRA, PEFT, transfer learning for domain fine-tuning. Built data engineering pipelines with Spark, Pandas, Databricks. Deployed & monitored production AI systems with CI/CD. Conducted hyperparameter tuning & experiment tracking for optimization. Researched Generative AI, contributing to open-source & internal innovation. Delivered solutions improving efficiency, reducing costs, and enhancing decisions.

At Zensar, I worked across classical ML and deep learning, building pipelines, deploying production models, and enabling data-driven decision making. Designed ML models for classification, regression, recommendation using scikit-learn, TensorFlow, PyTorch. Built data preprocessing pipelines for structured/unstructured data with Pandas, NumPy, Spark. Performed feature engineering & dimensionality reduction. Developed CNNs, RNNs, LSTMs for image recognition & NLP. Automated workflows with MLflow, Airflow, Docker. Deployed models via Flask APIs, FastAPI, Kubernetes on AWS. Applied cross-validation, GridSearchCV, Optuna, Ray Tune for tuning. Built real-time inference systems for anomaly detection & recommendations. Collaborated with data engineers to build ETL pipelines (BigQuery, Snowflake). Created dashboards & monitoring tools (Tableau, Grafana, Python scripts). Performed statistical analysis & hypothesis testing. Delivered REST APIs integrated into production microservices. Developed time-series forecasting (ARIMA, Prophet, LSTM). Used CI/CD (Git, Jenkins) for automation. Partnered with cross-functional teams to align ML solutions with business. Presented model insights & visualizations to stakeholders.