Sunita R

Working at the intersection of machine learning and financial risk, I design and deploy real-time ML systems that help trading and risk teams make faster, smarter decisions. I’ve built predictive models for portfolio risk (like Value-at-Risk and stress testing) using deep learning and ensemble methods, enabling teams to proactively manage exposure during volatile markets. I focus heavily on end-to-end ML systems—from feature engineering and data pipelines to deployment and monitoring—ensuring models are reliable, explainable, and production-ready. I’ve also contributed to fraud detection and insurance-related AI systems, improving accuracy while reducing manual effort for compliance teams. My work emphasizes scalability, low latency, and alignment with strict financial regulations.

I developed machine learning solutions for financial risk modeling and trading analytics, helping teams better understand and respond to market movements. My work included building time-series models and scalable pipelines that processed large volumes of market data in real time. I also focused on improving model reliability through automation—setting up workflows for retraining, monitoring drift, and ensuring consistent performance. Alongside finance, I worked on insurance AI use cases, evaluating model outputs for fairness and accuracy. A key part of my role was collaborating with business teams to translate complex requirements into practical ML solutions.

At McKinsey, I built NLP-powered solutions for clients across industries, helping them extract insights from large volumes of unstructured text. I worked on tasks like document classification, sentiment analysis, and entity recognition using transformer models and deep learning. I also optimized model performance and scalability, making sure solutions could run efficiently in production. My work included deploying APIs, improving training efficiency, and building explainable AI systems so stakeholders could trust model outputs. This role gave me strong exposure to client-facing problem solving and delivering business impact through AI.

I specialized in natural language processing solutions for healthcare and insurance domains. My work focused on extracting meaningful insights from clinical notes and insurance documents using techniques like NER, text classification, and sentiment analysis. I built end-to-end pipelines—from data cleaning and feature engineering to model deployment as APIs—helping automate processes like claims handling and document routing. I also worked closely with domain experts to ensure models aligned with real-world workflows and regulatory needs.

I started my career working on data analytics and machine learning models for insurance and business insights. I helped build data pipelines, clean large datasets, and develop predictive models for customer behavior, claims, and retention. I also created dashboards and reports to help stakeholders track performance and make data-driven decisions. This role built my foundation in data analysis, machine learning, and working with large-scale data systems.