Tejas Khandwekar

Tejas Khandwekar

Data Scientist & Analytics Professional

Atlanta, GA 30363
tejas5589@gmail.com
(404) 396-9196
LinkedIn
GitHub

Professional Summary

Experienced Data Scientist and Analytics Professional with 3+ years of expertise in machine learning, AI, and advanced analytics. Specialized in time series forecasting, natural language processing, and deep learning with proven track record of delivering $50M+ business impact. Currently pursuing MS Analytics at Georgia Tech while seeking Data Scientist, AI Scientist, and ML Engineer roles in time series analysis and NLP domains. Passionate about leveraging cutting-edge AI technologies to solve complex business challenges and drive data-driven decision making.

Education

Georgia Institute of Technology
Master of Science in Analytics
Georgia Institute of Technology
Atlanta, Georgia
August 2025 - December 2026
Bachelor of Technology in Mechanical Engineering
Visvesvaraya National Institute of Technology
Nagpur, India
July 2018 - July 2022

GPA: 9.21/10

Coursework: Machine Learning for Engineer, Numerical Methods, Machine Vision

Work Experience

Data Scientist
ExxonMobil
Bengaluru, Karnataka
July 2022 - June 2025
  • Automated fee calculation from digitized contracts using AWS Textract and developed a contract-invoice reconciliation system to detect mismatches using AWS Bedrock by Attribute Extraction using RAG, improving accuracy and efficiency in contract management.
  • Integrated GenAI-based fee calculator with AWS Glue ETL workflows, identifying invoice discrepancies and delivering $2.9M in cost savings.
  • Co-developed a global demand-forecasting pipeline for thousands of SKUs, addressing data gaps and tailoring ensemble models to segmented clusters, incorporating demand planners' feedback and business processes, and reduced model run times by 50%.
  • Improved demand-forecasting accuracy by 10% over vendor solutions, enabling $50M in business impact through enhanced planning and procurement.
  • Developed a statistical forecasting model for price predictions, enabling better capacity planning and trading decisions, achieving direction accuracy of 78%.
  • Achieved $2M/year financial savings by collaborating with domain experts, incorporating leading drivers, and using a custom loss function to minimize direction loss for trading.
  • Built a Dockerized segmentation model, which automated clustering, reduced analysis and workload time from months to hours, and boosted tailored marketing efforts, driving significant sales growth.
Project Engineering Intern
ExxonMobil
Bengaluru, Karnataka
May 2021 - July 2021
  • Created an information management tool using Power BI and SharePoint lists to organize deliverables and improve accessibility of information, storing and displaying data on a Project information portal written in TypeScript.

Skills

Programming & Tools

Python R SQL Power BI AWS Microsoft Office Suite

Machine Learning & AI

Machine Learning Deep Learning NLP Gen AI LLMs RAG

Analytics & Data Science

Time Series Analysis Demand Forecasting Statistical Analysis Customer Segmentation Anomaly Detection Supply Chain Analytics

Professional Skills

Analytical Thinking Strong Communication Collaborative Adaptable Version Control Actionable Insights

Academic Projects

iETS Models for Intermittent Demand Forecasting
Lancaster University
June 2024 - Present
  • Collaborated with Prof. Ivan from Lancaster University to test innovation state space models (iETS) for forecasting intermittent demand.
  • Applied R's smooth package to improve forecasts for irregular demand scenarios in retail sales data vs other models.
Natural Language Processing for Question Answering
Visvesvaraya National Institute of Technology
May 2021 - January 2022
  • Enhanced state-of-the-art NLP models for multi-hop question answering across multiple documents.
  • Built knowledge graphs and applied unsupervised methods for reasoning sentence selection; co-trained LSTM and fine-tuned BERT for improved QA performance.
Automated Vibration Comfort Classification
Visvesvaraya National Institute of Technology
Academic Project
  • Engineered a comprehensive vehicle vibration dataset using control theory in MATLAB, subsequently implementing 1D CNNs to accurately categorize time-series vibrations into comfort scores.
  • Effectively automated vibration classification and eliminated subjective human review through extensive feature engineering and model selection.
  • Built a custom version 1D ResNet (CNN) to classify time series data after analyzing different methodologies and models.
Porous Media Solver using CNN
Visvesvaraya National Institute of Technology
Academic Project
  • Developed a novel CNN model for simulating fluid flows in porous media, achieving a simulation speed approximately 2000 times faster than conventional CFD methods while maintaining legible results.
  • Pioneered the creation of a comprehensive dataset for this AI-driven approach, significantly contributing to the field and demonstrating the potential of machine learning in expediting innovation and design processes.

Achievements

Runner-Up, Myntra Hacker-Ramp Hackathon

Developed a CNN-based fashion trend detector using Instagram data.

ExxonMobil India Inc Recognition

Awarded for driving 10% accuracy gains in demand forecasting by Lead Country Manager.

Volunteering Recognition

Recognised for serving as the Recruitment coordinator for the department and fostering strong industry connections, resulting in successful recruitment outcomes at Visvesvaraya National Institute of Technology.

Additional Courses

Links available on LinkedIn

Machine Learning & AI

Mathematics for Machine Learning specialization (3 courses) – Deeplearning.AI
Supervised Machine Learning: Regression and Classification – Stanford Online
Interactive Machine Learning Dashboards using Plotly Dash – Coursera

Mathematics & Statistics

Linear Algebra from Elementary to Advanced (3 courses) – Johns Hopkins University
Statistics Foundations: The Basics – LinkedIn

Programming & Development

Object Oriented Programming with Python – LinkedIn
Git Essentials Training – LinkedIn

Project Management

Agile Development Practices – LinkedIn