Saibek Panicker

Hilton Analytics:

Working as a forecasting and planning analyst on the client's team for volume forecasting over each day.

• Enhanced accuracy by 10% by leveraging Machine Learning techniques, including FB Prophet, SARIMAX, and exponential smoothing, for advanced analytics tasks.
• Achieved over 90%+ forecast accuracy for statistical modeling & time series analysis across 19 groups
• Achieved a 60% reduction in forecasting code runtime for multiple groups by implementing Python's multiprocessing techniques.
• Facilitated cross-functional collaboration by optimizing staffing levels and integrating forecasts from Operations using Alteryx workflows.
• Implemented neural forecast methods to achieve better MAPE scores.

Latest Estimate forecast (LE Forecast):

The primary goal of this project is to forecast the revenue for Sutherland for the next 6 quarters. This project is a mix of classic ML scenarios of classification, regression, and it also includes time series forecasting as well. The major scenarios involved are:

1. Propensity to convert: To find the closed stage of an individual deal, i.e., to determine whether the deal is a won deal or a lost deal. A classification model is used for this purpose. Received an accuracy of over 92% for different scenarios.

2. Time to Convert: If a deal is a won deal, the next objective is to find out the duration taken for that deal to convert to the won. Inflows Forecast: The objective here is to forecast for the next six quarters on a vertical level. Applied extensive EDA, multiple machine learning, and forecasting models to obtain the results.

3. Inflows Forecast: The objective here is to forecast for the next six quarters on a vertical level. Applied extensive EDA, multiple machine learning, and forecasting models to obtain the results.

CX 360 Projects :

• ZENOTI: Worked on building and modifying multiple KPIs using various NLP techniques, particularly pattern matching, and obtained 90% accuracy for QA score.
• CUNA : Utilized GUSE & FAISS to implement NLP techniques, achieving over 90% accuracy in building a KPI for CUNA non-sales
• MARS: Worked on improving accuracies to above 90% for a few KPIs. Worked in auditing calls by agents to find out particular patterns for sentimental analysis.

• Led a team of four individuals to complete multiple projects ahead of schedule.
• Demonstrated expertise in regression analysis, time series forecasting, Convolutional Neural Networks (CNNs), and classification tasks.
• Developed a client project template for sales effectiveness, achieving an accuracy rate of over 90%.