Nelson Jyostna Wasekar

About Nelson Jyostna Wasekar

Experienced in developing and evaluating machine learning solutions for large-scale business applications. Delivered up to $29M in annual impact through a distributor-focused Recommendation System using GMM and NMF. Optimized waste management with a Random Forest model, reducing waste by 0.5% and saving $65K annually. Resolved token failures and enhanced Route Optimization with K-means clustering, boosting field efficiency. Built SVM-based deduction analytics to reduce $300K in revenue leakage. Led reverse knowledge transfer for Route Optimization and developed ML models in HyperScience OCR, increasing document processing efficiency by 20% through layout mismatch analysis and model structuring for diverse document types.

Experience

1. Software Engineer

   Tech Mahindra

   Jul 2022 — Present · Mumbai, IN

   Developed and evaluated a distributor-focused Recommendation System for Suggested Orders usingGMM and NMF clustering, enabling personalized up-sell, and innovation strategies and driving $1.5M-$2.4M monthly uplift, with an estimated annual business impact of $18M–$29M. Tracked input/output data streams and evaluated model performance for the YK Waste Management system using a Random Forest Classifier with 46 input features for TJ22 diaper production asset. Monitored Azure Event Hub message flow to ensure real-time data capture and system reliability & Azure Functions for orchestration and Azure Service Bus for stream processing. Working on a GenAI-powered PoC for automated ServiceNow ticket recommendations using LangChain, RAG pipelines, text embeddings, and ChromaDB vector database to enhance developer productivity. Involved in analyzing and addressing data leakage issues in the Peru baseline use case. Resolved token generation failures in production pipelines for the SFA and Chile regions of Route Optimization (RO) use case. Analyzed input/output data streams and assessed K-means clustering model performance of Brazil region, incorporating merchandisers’ weekly capacity and required hours to minimize travel distances. Monitored frequency and duration bucket creation, managed daily plan scheduling, and implemented dynamic cluster realignments to enhance operational efficiency and optimize field execution.Conducted reverse knowledge transfer and documented the As-Is and To-Be (AUD) process for the Route Optimization use case.Analyzed input/output data streams and evaluated the performance of a Deduction Analytics AI solution using historical data and Support Vector Machine (SVM) models to predict Level 2 reason codes for new deductions and classify them as valid or invalid.

Education

• Ramrao Adik Institute of Technology

  Master of Engineering (M.Eng.)

  2016 — 2019

• University of Mumbai

  Bachelor of Engineering - BE