I'm a data scientist interested in product analytics, experimentation, and LLM evaluation.

I turn messy, raw data into decisions — building predictive models, data pipelines, and the ML and LLM systems that put them to work.

Journey so far
Six years, six moments that shaped the judgement
Milestone
Coursework
Fall 2020
B.S. in Computer Science
Ahmedabad University
Started my undergrad, where I first got hooked on data and machine learning.
Summer 2023
Data Engineering Intern
Travel Designer Group Pvt. Ltd.
Built an automated ETL pipeline (ELK stack) giving real-time visibility into 1M+ booking searches.
Spring 2024
Data Science Intern
Softvan Pvt. Ltd.
Applied statistical analysis and text mining to 50,000+ customer records, driving a 15% lift in retention.
Fall 2024
M.S. in Data Science
UT Arlington
Moved to Texas to specialize in machine learning, big data systems, and data visualization.
Mid-Spring – Fall 2025
Student Data Analyst
UT Arlington
Automated a Power BI dashboard over 150,000+ CRM records, saving 80+ hours of manual work monthly.
Spring 2026
Graduate Data Science Intern
Delta Air Lines
Cut document processing time 12x with an AWS SageMaker NLP pipeline; lifted GenAI summary quality from 3.5 to 4.2/5.
Fall 2020
Programming, Mathematics
Spring 2021
Object Oriented Programming
Summer 2021
Volunteering at an animal welfare trust
Fall 2021
Data Structures, Applied Linear Algebra
Spring 2022
Database Management Systems, Probability & Stochastic Processes
Fall 2022
Probabilistic Graphical Models
Spring 2023
Big Data Analytics, Machine Learning
Fall 2023
Artificial Intelligence, Predictive Analytics for Business
Fall 2024
Python, Data Science, Probability & Statistics
Spring 2025
Machine Learning, Data Visualization, Big Data Management
Fall 2025
Project Management, Applied Regression Analysis
2020
2021
2022
2023
2024
2025
2026
View my work Get in touch

Experience

Download PDF
Jan 2026 — May 2026
Atlanta, GA
Graduate Data Science Intern
Delta Air Lines
  • Cut document processing time 12x by engineering an end-to-end AWS SageMaker pipeline applying NLP to 7,000+ unstructured maintenance/ops PDFs.
  • Improved generative AI summary quality from 3.5 to 4.2/5 by running before/after scoring sessions with station managers and optimizing prompts.
  • Turned a 5-minute read into under 30 seconds for 1,000+ users.
Mar 2025 — Dec 2025
Arlington, TX
Student Data Analyst
University of Texas at Arlington
  • Saved 80+ hours of manual work monthly with an automated Power BI dashboard pulling 150,000+ CRM records.
  • Unified calls, emails, and walk-ins into one operational view, giving leaders the insight to optimize staffing and cut wait times at peak hours.
Jan 2024 — May 2024
Ahmedabad, India
Data Science Intern
Softvan Pvt. Ltd.
  • Applied statistical analysis and text mining to 50,000+ customer records and survey responses to surface behavioral segments, driving a 15% improvement in retention.
  • Presented findings to non-technical stakeholders through clear Tableau dashboards.
Jun 2023 — Aug 2023
Ahmedabad, India
Data Engineering Intern
Travel Designer Group Pvt. Ltd.
  • Gave the business real-time visibility into 1M+ booking searches by building an automated ETL pipeline (ELK stack) pulling live data through APIs.
  • Designed Kibana dashboards breaking down searches by location and event type to spot booking trends.
Education
Aug 2024 — May 2026
Arlington, TX
M.S. in Data Science, University of Texas at Arlington
Aug 2020 — May 2024
Ahmedabad, India
B.S. in Computer Science, Ahmedabad University
Skills
Languages
PythonSQLR
Databases
PostgreSQLpgvectorPinecone
Machine Learning & Statistics
RegressionClassificationClusteringTime-Series ForecastingA/B TestingHypothesis TestingFeature EngineeringModel Training, Selection & Evaluation
Libraries
Scikit-learnXGBoostPyTorchTensorFlowPandasNumPystatsmodels
Big Data & Cloud
Spark (PySpark)DatabricksAWS (S3, SageMaker, Athena, EC2)Azure
MLOps & Deployment
MLflowDockerCI/CD (GitHub Actions)FastAPI
GenAI & LLM
Document ProcessingStructured Output ExtractionAgentic WorkflowsTool CallingMCPRAGPrompt EngineeringLLM Evaluation
Libraries
LangChainLangGraphHugging FaceDocling
Visualization & BI
Power BITableauKibanaMatplotlib
Version Control
GitGitLab

Projects

2025 – 2026
NutriBot
Agentic RAG nutrition assistant
LangGraphPineconeFastAPIAWS
  • LLM-controlled retrieval routes across multiple sources
  • Self-grading loop drops low-relevance context before answering
  • Result — 88% grounded-answer rate on a held-out question set (RAGAS faithfulness), with the self-grading step cutting hallucinated responses
View code / demo →
DelayCast
Flight-delay prediction at scale
PySparkXGBoostMLflowDatabricks
  • Engineered features from large flight datasets in PySpark
  • Binary classifier tracked and versioned with MLflow on AzureML
  • Result — 0.73 ROC-AUC on held-out flights, outperforming the majority-class baseline
View code / demo →
Electricity demand forecasting
Short-term load forecasting
PythonXGBoostscikit-learn
  • Lag and calendar features capture daily and seasonal demand patterns
  • Forecasts demand over a 12 step horizon
  • Result — 5.71% MAPE on next-day demand, improving on a seasonal-naïve baseline
View code / demo →
Loan default prediction
Credit-risk default classifier
PythonXGBoostscikit-learn
  • Handled class imbalance so rare defaults aren't ignored
  • Probability-calibrated scoring for usable risk thresholds
  • Result — 0.93 F1 on the minority (default) class after imbalance handling
View code / demo →

About

I care about the unglamorous middle of data science — the cleaning, the pipelines, the sanity checks — because that's where messy, real-world data becomes something a team can actually decide with. Across an airline, a university, and early-stage startups, I've built forecasting models, ML pipelines, and RAG systems, always with an eye on whether the output holds up in production.

I'm completing my M.S. in Data Science at UT Arlington (May 2026), following a B.S. in Computer Science at Ahmedabad University. Most recently I was a Graduate Data Science Intern at Delta Air Lines in Atlanta, building NLP pipelines for aircraft operations. I'm now looking for full-time roles where I can own problems end to end.

Let's talk data.

Open to new roles and interesting problems. Reach out anytime.