Open to opportunities

VAIBHAV KHANDELWAL

@vaibhavkhandelwal

Message

I build enterprise AI automation, RAG, forecasting, and machine learning systems that deliver measurable savings.

India

Message

What I'm looking for

I'm looking to build secure, scalable AI automation and machine learning solutions that connect LLMs, RAG, forecasting, and enterprise systems to deliver measurable operational efficiency and business value.

I've built AI and machine learning systems for Allstate, Sony Europe, Forsberg Ltd., and Infosys, delivering automation, forecasting, and decision-support solutions with measurable business impact.

At Allstate, I led the architecture of a Hybrid RAG automation system that classified service-call transcripts across 65 intents at 90%+ accuracy using AWS Bedrock, Lambda, and Jenkins. I also delivered a property insurance loss prediction model that generated $5.5 million in realized bottom-line savings.

At Sony Europe, I productionized logistics optimization and hierarchical time-series forecasting for more than 1,000 dealers, reducing operational expenditure by 30%. I also built Azure Databricks NLP sentiment-analysis platforms using BERT and GPT, alongside REST API integration and data-governance workflows that automated approvals by 95%.

I bring hands-on experience across AWS, Azure, Python, SQL, PySpark, LLMs, RAG, agentic workflows, and secure microservices. I enjoy translating complex business problems into scalable AI architectures, practical automation, and reliable production systems.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist Consultant

Current

Allstate

Jan 2025 - Present (1 year 7 months)

Led end-to-end architecture of a Hybrid RAG automation system to classify service call transcripts into 65 intents with 90%+ accuracy; productionized using AWS Bedrock, Lambda, and Jenkins. Architected an automated Theme Identification framework leveraging AWS strands and multi-agent orchestration to classify service transcripts with high precision.

Lambda Jenkins Hybrid RAG Multi agent orchestration Pydantic GuardRails Microservices Sagemaker

Data Scientist

Sony Europe

Jan 2023 - Oct 2023 (9 months)

Prototyped and productionized complex greedy optimization algorithms and multi-layered data ingestion services to automate global logistics tracking loops. Architected a scalable hierarchical time-series forecasting engine to optimize supply chain and shipping logistics for 1,000+ dealers; improved demand planning accuracy by integrating Croston-based models, resulting in a 30% reduction in operat

Time Series Modeling Croston's method REST APIs Collibra Databricks Azure BERT GPT NLP Sentiment Analysis

Data Scientist

Forsberg Ltd.

Jun 2022 - Sep 2022 (3 months)

Developed an ETL pipeline for satellite data anomaly detection on Azure, utilizing the advanced statistical Recursive Density Estimation(RDE) method, achieving a 95% accuracy in detecting jamming. Crafted an insightful Tableau dashboard for analyzing points of jamming, delivering actionable business insights to key stakeholders, and enhancing decision-making capabilities.

ETL Azure Recursive Density Estimation Tableau Anomaly Detection

Data Scientist

Infosys

Jan 2018 - Sep 2021 (3 years 8 months)

Developed and executed time-series based forecasting algorithms to forecast future tax collection for a taxation project consisting of 150 million businesses across the country. Developed an unsupervised cluster-based Outlier Analysis to detect fraudulent entities by analyzing clusters using supply-based ratios, considering business knowledge and the silhouette method to determine the optimal numb

Time Series Analysis Unsupervised Learning Clustering Outlier Analysis PySpark Hive SQL Tableau