SB

Soumya Baddham

Available to hire

Hi, I’m Soumya B Addham, a data scientist with 4+ years of experience in machine learning, NLP, time-series forecasting, and GenAI.

About

Hi, I’m Soumya B Addham, a data scientist with 4+ years of experience in machine learning, NLP, time-series forecasting, and GenAI. I enjoy turning complex data into practical solutions that empower teams and drive business outcomes.\n\nI’ve built production pipelines, NLP models, and MLOps workflows, and I love delivering insightful dashboards that make data accessible across stakeholders. I’m open to new opportunities and relocating as needed.

  • 4 years' experience
  • Joined December 2025

Languages

  • EnglishFluent

Industry experience

  • Financial Services
  • Professional Services

Work experience

Data Scientist · Morgan Stanley USANov 2024 – Present · 2 yrs 1 mo

Led the development of large-scale financial data pipelines (PySpark, AWS Glue, and Athena), restructuring ETL flows and enforcing schema validation to improve intraday data availability by 38% and meet regulatory data quality standards. Built a production-ready Retrieval-Augmented Generation (RAG) workflow with LangChain and GPT-4, reducing analyst research time by 55% and increasing the accuracy of investment insights through enhanced retrieval precision. Developed transformer-based NLP models (RoBERTa) and Named Entity Recognition (NER) to extract counterparty, transaction, and risk-related entities from unstructured reports, boosting extraction accuracy from 82% to 94%. Implemented real-time trade-surveillance analytics using AWS Kinesis and TensorFlow, enabling proactive anomaly detection and reducing false alerts by 22%. Strengthened MLOps with model versioning, drift checks, and CI/CD pipelines, reducing release cycles by 40% while ensuring regulatory compliance. Produced intera

Jr. Data Scientist · Zensar Technologies IndiaJan 2020 – Dec 2022 · 3 yrs

Built supervised ML models (Decision Trees, Random Forests, SVM, Naive Bayes, XGBoost) to predict customer churn and support-ticket patterns, achieving up to 92% accuracy and enabling early retention interventions. Developed time-series forecasting models using ARIMA and Facebook Prophet to project weekly demand, reducing forecast error by 27%. Led Exploratory Data Analysis (EDA) and hypothesis testing on 10M+ transactional records to identify revenue leakage and process gaps, contributing to an 18% improvement in operational efficiency. Created CNN-based image classification pipelines using Keras and PyTorch to achieve 96% precision on defect-detection tasks for a manufacturing client, reducing manual quality checks by 40%. Automated ETL pipelines in Python (NumPy, Pandas) for weekly operations, reducing processing time from 3 hours to 20 minutes, and designed Power BI dashboards to cut manual reporting time by 50%.

Education

Master in Computer Science · University of Kansas2030 – 2025

Hire Soumya Baddham today

Post a job and invite Soumya, or contact them directly. No hidden commission.

Post a job