Divyank R

Implemented ML-based multi-task project cost models that increased estimation accuracy and slashed estimation time. Developed and deployed Azure web apps with Azure DevOps CI/CD pipelines, enabling project managers to instantly run cost prediction models and accelerating decision-making. Engineered a Random Forest Regression model achieving 95% accuracy in forecasting FTE costs, improving budget allocation and margin calculation across multiple construction projects. Led an end-to-end ML proof-of-concept (POC) by creating a Python Flask frontend and integrating advanced topic modeling (BERT, TF-IDF) for internal project document classification, demonstrating complete solution ownership from model to user interface. Built a multi-stage NLP pipeline for survey data (10,000+ responses) with 92% classification accuracy and 88% F1-score in sentiment analysis, directly informing new engagement policies for 5,000+ employees.

(NASDAQ: NEGG) Built a product recommendation engine, lifting Average Order Value by 18% and adding $30M in annual revenue. Designed and deployed inventory optimization algorithms that balanced stock levels across distribution centers, improving fulfillment efficiency by 22% Led 100+ A/B tests, improving conversion rate by 7% (~$15M incremental sales). Performed customer segmentation with Spark ML, identifying segments generating 45% of revenue and elevating retention by 25%. Orchestrated LSTM forecasting and GCP pipelines (BigQuery, Dataflow), slashing inventory costs by 15% (~$8M) and stockouts by 35%. Developed real-time fraud detection using ensemble methods, cutting fraud by 60% and saving $12M yearly.

Compared different deep learning models like CNN, VGG16, Inception Net and RESNET on ASL database. Evaluated the model efficacy using accuracy parameter on the test data.

Automated data processing workflows using Python and Excel, integrating structured and unstructured data into Tableau dashboards, reducing reporting time by 70%. Collaborated with cross-functional teams to standardize data quality and promote data-driven decision-making across multiple departments.