Open to opportunities

Somita Chaudhari

@somitachaudhari

Data Scientist with 5 years delivering production ML, forecasting, segmentation, and LLM pipelines for measurable business impact.

United States

Message

What I'm looking for

I’m looking for a team where I can build and deploy production ML and LLM pipelines—owning forecasting, segmentation, and document intelligence end-to-end—while partnering closely with data engineering and business stakeholders to deliver measurable impact.

I’m a Data Scientist with 5 years of experience deploying production-grade machine learning and deep learning solutions across supply chain, e-commerce, and healthcare—driving $7.8M in measurable business impact. I specialize in demand forecasting, customer segmentation, churn prediction, and inventory optimization, with a strong focus on building reliable pipelines and model lifecycle practices.

In my most recent role at Merkle, I reduced excess inventory costs by $800K through an AWS-based demand segmentation model with drift monitoring and automated retraining, and I built an LLM-based document intelligence pipeline that cut manual review effort by 35% for a Fortune 500 retail client. Previously at LTIMindtree, I led end-to-end DS engagements and delivered a demand forecasting system that saved $5M in excess inventory and stockout-related losses, while also contributing a $2M revenue uplift through forecasting models. I bring collaborative, stakeholder-driven problem solving—scoping KPIs, aligning features and schemas across systems, and shipping solutions that teams can operationalize.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist

Current

Merkle

Jul 2025 - Present (10 months)

Built a demand segmentation model on AWS to reduce excess inventory costs by $800K and supported supply planning with 5TB behavioral feature store drift monitoring and automated retraining. Delivered an LLM-based document intelligence pipeline that reduced manual procurement/vendor contract review effort by 35% for a Fortune 500 retail client.

Demand Forecasting Customer Segmentation Inventory Optimization LightGBM AWS S3 AWS Sagemaker Drift Monitoring

Data Analyst

University of Maryland, Baltimore County

Aug 2024 - May 2025 (9 months)

Built automated ETL pipelines on AWS to reduce manual data preparation time from 2 weeks to 3 days by integrating enrollments, usage logs, and e-commerce orders. Created Tableau dashboards and analytics on 150K+ customers, improving subscription conversions by 15%.

ETL Python PostgreSQL AWS Lambda AWS RDS Tableau A B Testing XGBoost Data Visualization

Data Scientist

LTIMindtree

Jul 2020 - Jul 2023 (3 years)

Owned end-to-end data science engagements for two Fortune 500 retail clients, delivering demand forecasting and inventory optimization models to drive measurable business outcomes. Engineered revenue forecasting models contributing $2M uplift in revenue and built SKU-level demand forecasts that saved $5M in excess inventory and stockout-related losses.

Demand Forecasting Revenue Forecasting Inventory Optimization XGBoost Pyspark AWS Sagemaker Tableau

Data Science Intern

LTIMindtree

Jan 2020 - Jun 2020 (5 months)

Performed HIPAA-compliant data wrangling in SQL on EHR records by cleaning, joining 5+ source tables, and standardizing patient utilization data for equipment inventory analytics. Supported senior data scientists with feature engineering, exploratory data analysis in Python/SQL (window functions), and Tableau views for predictive model development.

SQL Python Feature Engineering Window Functions EHR Data Integration Tableau Data Modeling