Brajesh Kumar Jha

• Currently spearheading the organizational Al transformation project with 6 Al Engineers
• Built a conversational licensing chat bot using Azure Al & Azure search service RAG architecture. Improved productivity by 8%
• Built a conversational chat bot on structured data (Text - to- Python) RAG architecture. Improved productivity by 3%
• Built a conversational chat bot on Offline contracts using Lang-chain RAG Architecture
• Built in building Agentic Al solutions for Marketing assist team using Crew Al that uses GPT- 4o
• Partnered with c-level stakeholders across Sales, Marketing, Operations to understand the pain points and formulate the execution plan
• AI solution automatically accept/ reject a partner application on the application merits. Classification model based on NLP and traditional ML
• Experience in showcasing progress and present in front of ELT (Executive leadership team) in story board format.

• Leading, Guiding and mentoring a team of 10 ML Engineers, AI engineers, Data Analyst
• Developed 1st year renewal model for multiple year policies from scratch. Used Light GBM method. This model calculates probability of default on renewal premium by customer before 90 days of premium due date. Leading to increase in customer retention by 3%
• Developed 2nd year renewal model for multiple year policies from scratch used Ensemble method. This model calculates probability of default on renewal premium by customer before 90 days of premium due date. Leading to increase in customer retention by 2%.
• Early Claim Prediction model at the time of issuance of new policy of claims. XGBOOST algorithm. Used Geospatial analysis of the claims and frauds.
• Developed Fraud detection model to flag the probable frauds at the time of claim intimation
• Data Analytics, Insights generation, Development of Prediction model for top performing field level executive. This model is deployed by the HR for recruitments

• Leading, Guiding and mentoring a team of 4 ML Engineers, Data Scientist, Data Analyst
• Development of Fraud detection model from scratch used Logistic Regression, Decision Tree, Random forest. This model calculates fraud score in the whole lifecycle of the claim- Claim Intimation, First Contact and till the time of Claim Closure. 2 separate model are developed for Motor and Health Claims respectively
• Recommendation system for Insurance products: Used data analytics to provide solution for next best insurance products for customers Used techniques like Collaborative filtering, SVD, KNN
• Claim Amount Prediction at the time First intimation. Used Image detection to calculate the amount of damage. Deep learning model, CNN model
• Churn Prediction with Ensemble and NLP Techniques: Built an ensemble-based churn prediction model using a hybrid of XGBoost, random forest, and text-based neural networks to analyze structured data and NLP-driven insights from customer interactions
• Leveraged sentiment analysis on customer support interactions as features, achieving an 18% improvement in churn prediction.

• Voltas Limited-Automated Service Request Classification with NLP: Built a NLP-based multi-class model to categorize service requests (e.g., maintenance, installation) with a 92% F1-score, cutting manual routing time by 60%. Deployed via REST API for real-time classification, enabling faster resolutions. Primarily this company is consumer electronics and engineering industry, specializing in air conditioning, refrigeration, and climate control solutions
• Customer Profiling and Revenue Analytics at Daimler India: Developed targeted marketing strategies through K-means clustering for customer segmentation and ARIMA forecasting, achieving a MAPE of 8%. Built decision trees for criticality assessment, enhancing engagement for high-value clients and improving sales conversion rates by 25%. Created real-time dashboards in Tableau
• Worked for IFFCO Tokio General Insurance Ltd.(as described above).

• Designing Prototype Design for Airline E-commerce: Designed a prototype for an airline e-commerce platform, leveraging user experience (UX) design principles and A/B testing to optimize customer journey. Implemented recommendation systems using collaborative filtering, enhancing upsell opportunities and increasing conversion rates by 18%
• Revenue Analysis and Forecasting: Utilized Python and SQL to build a revenue analysis model for e-commerce, applying time series analysis and linear regression to achieve 90% forecasting accuracy. Developed a dynamic pricing algorithm using machine learning techniques, resulting in a 15% revenue increase
• Business Intelligence Dashboard: Created an interactive dashboard in MS Office to visualize e-commerce KPIs, employing predictive modeling to uncover customer trends. Integrated data insights into financial planning, enhancing forecasting accuracy by 20% through advanced analytics techniques.

·Worked in Process Optimization, Cost Reduction, Battery Heating Process, Water Use.