AI/ML Engineer at Northern Trust, OH

Designed and deployed machine learning pipelines on AWS SageMaker, reducing model training time by 35%. Built predictive credit risk models using logistic regression and gradient boosting, improving risk assessment accuracy. Integrated NLP models for document classification and sentiment analysis, automating manual reviews and saving over 500 work hours annually. Collaborated with DevOps teams to implement MLOps workflows using Docker, Kubernetes, and GitHub Actions for automated deployments. Developed real-time data ingestion and preprocessing pipelines with Apache Kafka and Spark for scalable AI solutions. Designed a sentiment analysis pipeline leveraging NLP and Transformers, achieving 90% accuracy. Enhanced chatbot capabilities using Transformers, LangChain, LLaMA Index, and vector stores (ChromaDB, FAISS) to provide human-like responses. Conducted adversarial attack simulations (Prompt Injection, TAP, PAIR) resulting in a 15% improvement in AI model security.

ML Engineer at Aplus Datalytics, India

Designed end-to-end ML workflows from data ingestion to production deployment, managing datasets exceeding 100 million records. Applied supervised learning models (Random Forest, XGBoost, SVM) to improve fraud detection accuracy by 22%. Built unsupervised clustering models for customer segmentation, enabling targeted marketing that boosted engagement by 15%. Created time series forecasting for sales prediction, reducing inventory shortages by 20%. Automated feature engineering pipelines in Python, cutting data preparation time by 40%. Deployed AI models via REST APIs using Flask and FastAPI. Led knowledge transfer and mentored junior data scientists to improve team productivity. Developed a computer vision system with OpenCV and TensorFlow, automating quality inspection with 95% accuracy. Created ML pipelines for predictive analytics, improving business forecasting by 20%. Conducted feature engineering and model optimization to enhance model performance by 30%. Designed unit and integr

AI/ML Engineer at Northern Trust

Designed and deployed ML pipelines on AWS SageMaker, reducing model training time by 35%. Built predictive credit risk models using logistic regression and gradient boosting to improve risk assessment accuracy. Integrated NLP models for document classification and sentiment analysis, automating manual reviews and saving 500+ work hours annually. Collaborated with DevOps to implement MLOps workflows using Docker, Kubernetes, and GitHub Actions for automated deployments. Developed real-time data ingestion and preprocessing pipelines with Apache Kafka and Spark for scalable AI solutions. Designed a sentiment analysis pipeline leveraging NLP and Transformers, achieving 90% accuracy in sentiment classification. Enhanced chatbot capabilities using Transformers, LangChain, LLaMA Index, and vector stores (ChromaDB, FAISS) for more human-like responses. Conducted adversarial attack simulations (Prompt Injection, TAP, PAIR) to improve model security by 15%.

ML Engineer at Aplus Datalytics

Designed end-to-end ML workflows from data ingestion to production deployment for datasets exceeding 100M records. Applied supervised learning techniques (Random Forest, XGBoost, SVM) to improve fraud detection accuracy by 22%. Built unsupervised clustering models to segment customer profiles, enabling targeted marketing campaigns and boosting engagement by 15%. Created time series forecasting models for sales prediction, reducing inventory shortages by 20%. Automated feature engineering pipelines in Python (Pandas, NumPy, Scikit-learn), cutting data preparation time by 40%. Deployed AI models via REST APIs using Flask and FastAPI for client integration. Led knowledge transfer sessions and mentored junior data scientists. Developed a computer vision system using OpenCV and TensorFlow, achieving 95% accuracy in quality inspection. Created an end-to-end ML pipeline for predictive analytics, leading to a 20% improvement in business forecasting. Conducted feature engineering and model opti

AI/ML Engineer at Northern Trust

Delivered end-to-end ML pipelines on AWS SageMaker with MLflow tracking and registry; standardized reproducible experiments and approvals, cutting model training/packaging time by 35% and tightening governance for regulator reviews. Productionized NLP for document classification and sentiment routing; automated manual review queues, saving hundreds of analyst hours annually and reducing turnaround time. Shipped containerized inference with Docker/Kubernetes and FastAPI/Triton, implemented blue-green deploys and autoscaling; reduced P95 latency and infra spend per 1k requests. Built Kafka/Spark streaming ingestion and feature pipelines with data contracts and quality checks; improved feature freshness and reduced data incidents. Enhanced LLM-based chatbot workflows with LangChain, LlamaIndex, FAISS/Chroma; added safety guardrails. Introduced adversarial testing and drift/bias monitoring; strengthened AI security and governance. Collaborated with DevOps to codify CI/CD for data/feature/m

ML Engineer at Aplus Datalytics

Owned full ML lifecycle from ingestion to deployment across datasets >100M rows; standardized data validation, feature engineering, and model delivery templates for repeatability. Improved fraud detection with Random Forest/XGBoost/SVM and cost-sensitive learning; boosted precision/recall while keeping analyst load manageable via calibrated thresholds. Built unsupervised segmentation (K-Means/DBSCAN) to drive targeted campaigns; partnered with marketing to translate clusters into actions, lifting engagement by double digits. Developed time-series forecasts for sales/inventory using feature-rich regressors; reduced stock-outs and carrying costs through demand-sensing dashboards. Exposed models via REST (Flask/FastAPI) with request/response schemas, RBAC, and usage analytics; simplified integration for client apps and support teams. Mentored juniors and implemented code reviews, unit/integration tests (including data tests) to improve delivery velocity and quality. Built a CV inspection

AI/ML Engineer at Northern Trust

Designed and deployed deep learning-based fraud detection models using TensorFlow and PyTorch, improving anomaly detection precision by 22% across high-volume transactions. Architected and automated data ingestion pipelines on AWS (Glue, Lambda, S3) to handle 15M+ daily records with near real-time data availability. Applied NLP and Graph Neural Networks (GNNs) to detect relationship-based fraud and pattern anomalies in transaction graphs. Integrated AI models into production via AWS SageMaker Endpoints and Dockerized APIs, reducing latency for scoring events to sub-300ms. Implemented MLOps frameworks using MLflow, Airflow, and Jenkins for experiment tracking, version control, and automated retraining workflows. Partnered with Compliance and Model Risk teams to implement model explainability (SHAP, LIME) and AI governance standards. Created real-time monitoring dashboards in Grafana and Power BI to track fraud KPIs, precision drift, and operational performance metrics.