Bibhuti Bhusan Sahoo

WFI Planning & Estimation: Client - Allianz Insurance

Responsibilities:

• Estimating long-term and short-term forecasts for 250+ processes, providing actionable insights.
• Identifying anomalies and implementing remediation strategies.
• Evaluating trends, seasonality, cyclic patterns, and irregularities using statistical analytics tools.
• Automating forecasting pipelines to enhance efficiency and accuracy.
• Publishing strategic plans and guiding decision-making across the organization.
• Evaluating model performance using drift metrics such as KL Divergence, JS Divergence, Wasserstein Distance, PSI Index, and the Page-Hinkley Test, etc.
• Developing data-driven solutions to optimize business performance.

TPAIRS: Client - Citi Bank

Responsibilities:

• Worked alongside team members and leaders to identify analytical requirements and collect information to meet customer and project demands.
• Produced in-depth analyses and proposed solutions to meet complex business needs.
• Implemented ML algorithms to evaluate and solve diverse company problems
• Utilised data to enhance business development, marketing and other areas
• Reviewed business areas to strengthen analytics, enhance strategic planning and support decision-making processes with solid data
• Used advanced analytical approaches to evaluate transaction and department key factors, including the impact on company profits and growth potential.

VIVA (Virtual Intelligence Voice Analyser):

Project Objective:
Develop an automated VoiceID verification system leveraging Signal Processing and advanced Machine Learning algorithms to distinguish fraudulent and genuine calls from flagged audio recordings in HSBC’s Contact Centre.

Responsibilities:

• Analyzed large-scale voice call datasets, programmatically segmenting individual calls to extract customer-specific audio from multi-channel recordings.
• Extracted key audio features, including SNR, Energy, Spectral Features, MFCC, and pitch features, from various samples.
• Developed classification models using Machine Learning and Deep Learning techniques with Scikit-Learn, Keras, and TensorFlow.
• Designed an end-to-end project framework, integrating the model with a front-end solution and ensuring seamless deployment.
• Successfully reduced operational fraud losses by £1 million per month, post-implementation, in the Fraud Contact Centre.

CCS-Gini:

Project Objective:
An automated business assistive system designed to provide customers with end-to-end visibility into their application journey, including estimated TAT, application status, queue-level TAT, approval timelines, decline reasons, mean wait hours, and more. Additionally, it supports operations teams by offering insights into application volumes, agent RPH, priority application status, cancellation reasons, and other key performance metrics.

Responsibilities:

• Gathered and summarized requirements, assigning task modules to the team for efficient development.
• Developed predictive models for estimating TAT, mean wait hours, and other key application metrics.
• Built modules for Natural Language Generation (NLG) and Natural Language Understanding (NLU) to enhance system interactivity.
• Supervised UI integration planning to ensure seamless functionality and user experience.

U5-Capacity Planning:

Project Objective:

Project Objective: Developed a forecasting system to predict daily staff demand for verifying customer applications across HSBC’s operational floors for various sub-products, including cards, loans, mortgages, credit limit increases (CLI), and overdrafts. The project was globally deployed across all HSBC banking regions.

Responsibilities:

• Gained a deep understanding of operational processes to align forecasting models with business needs.
• Forecasted TTD (Turnaround Time for Decision) and referred application volumes using Time Series Modeling techniques.
• Conducted shrinkage and productivity analysis at both product and country levels, developing predictive models to estimate staff demand based on multiple factors.
• Presented reports and insights in governance and RTB (Run-the-Bank) forums, collaborating with stakeholders for strategic decision-making.

• Intend of the project to develop supervised models for closing price of nearly 6000 stock tickers available under NYSE and NASDAQ and perform intraday prediction for all stocks
• These models will help customer portfolios to study future behaviours/trends individually and benefits them in terms of monetary aspects
• Extracted last 20yrs of daily historical data for all stock tickers, enlarged predictors by including performance indicators and beta features
• Developed individual models for all tickers and integrated with front end
• Created automated script which will predict intraday closing prices before Stock Exchange starts
• Intend of the Project for Cluster analysis of call handle time vs
• Various factors such as skill, campaign, customer gender, geographic areas, age, agent, background etc
• Predicting Resources needed in order to improve overall productivity as well as Wrap up Time, Queue Time, Hold Time, etc
• Analysing Resource & Agent Reliability by applying Predictive Modelling to optimise cost of the organisation

• Intend of the project is to understand the impact of climate change on cryospheric contribution to rivers in western Himalaya and the subsequent impact on water resource availability and socio- economic status of the local community
• Analysing IMD temperature and rainfall historical data, understanding polar climatic changes, monthly and seasonal analysis at multiple IMDB stations
• Predicting temperature and rainfall over research regions by identifying significance predictors using feature selection methods and presenting the analysis to stakeholders