Parth Arora

About Parth Arora

As a Data Scientist with over 3 years of experience, I specialize in building end-to-end machine learning and AI-driven systems that solve real-world business problems. My core expertise spans predictive modeling, deep learning, NLP, and statistical experimentation, with a proven ability to transform complex datasets into strategic insights. I’m passionate about delivering scalable AI solutions using tools like Python, SQL, TensorFlow, PyTorch, and leveraging cloud platforms such as AWS and Azure for robust deployment. Currently at Goldman Sachs, I design and deploy ML systems to optimize financial risk modeling, client segmentation, and regulatory compliance. My work includes deploying real-time deep learning models for fraud detection, fine-tuning LLMs like GPT and Claude for automation, and building scalable pipelines for NLP tasks such as sentiment analysis and document summarization. I also lead efforts in MLOps by integrating CI/CD, MLflow, Docker, and Kubernetes into production-ready workflows. Before Goldman Sachs, I honed my skills at LTIMindTree, where I developed data-driven solutions across BFSI and retail, automated ETL pipelines, deployed forecasting and churn prediction models, and presented business-aligned insights using Tableau and Power BI. My experience also includes working closely with cross-functional teams to ensure alignment between technical outputs and organizational goals. I’m driven by curiosity, impact, and continuous learning. Whether it’s developing real-time AI products, building intelligent assistants with LLMs, or experimenting with causal inference and generative models, I’m excited by the intersection of data and innovation. I\'m always open to new opportunities that let me apply cutting-edge AI to solve meaningful challenges and create scalable, user-centric solutions.

Experience

1. Data Scientist

   Goldman Sachs

   Nov 2023 — Present · Jersey City, NJ, US

   Designed and deployed machine learning models to enhance financial risk modeling and investment strategies, integrating over 120 million data points from structured and unstructured sources, resulting in a 22% increase in predictive accuracy across portfolio simulations.2. Constructed deep learning pipelines using TensorFlow and PyTorch to detect fraudulent activity and credit risk in real-time transactional datasets, lowering false positives by 28% and accelerating regulatory alerting by over 30%.3. Built and activated end-to-end NLP solutions using Hugging Face Transformers and spaCy to analyze over 1.5 million financial documents annually, delivering sentiment analysis, entity recognition, and classification models with an 88%+ precision rate.4. Fine-tuned large language models (LLMs) like GPT and Claude for internal automation use cases including investment memo drafting, risk commentary summarization, and policy adherence checks, leading to a 42% reduction in manual documentation overhead across analyst teams.5. Developed reusable prompt templates and evaluation frameworks for LLM output quality control, enabling consistent generation of client-facing content, internal FAQs, and regulatory summaries while minimizing review cycles by 37%.6. Delivered tailored Tableau dashboards to visualize model outcomes, market indicators, and NLP-derived insights in real time, empowering senior leadership with actionable intelligence and reducing quarterly reporting turnaround by nearly 25%.7. Built end-to-end model monitoring pipelines using MLflow, custom validation scripts, and CI/CD integration to ensure traceable experiments, continuous drift detection, and seamless retraining of over 15 live models, maintaining 99.3% uptime and full audit readiness across production systems.

Education

• Stevens Institute of Technology

  Master of Science - MS

  2022 — 2024

• Maharaja Surajmal Institute Of Technology

  Bachelor of Technology - BTech

  2018 — 2022