🎓 M.S. in Business Analytics & Artificial Intelligence @ UT Dallas
📍 Richardson, Texas, USA
💡 Turning messy data into intelligent, production-ready systems
I love building end-to-end data & AI systems that go beyond Jupyter notebooks:
- ✅ Real ETL & ELT pipelines (Spark, DuckDB, AWS)
- ✅ Analytics-ready datasets with proper modeling & governance
- ✅ Dashboards & insights that non-technical stakeholders can trust
- ✅ Gradually moving into MLOps & model monitoring
My background in Electronics & Communications Engineering (VIT) + my Master's at UT Dallas gives me a nice mix of math + engineering + business.
| Languages | ML & Data | Data Engineering | Cloud & BI |
| Python · R · SQL · HTML/CSS | NumPy · Pandas · Scikit-Learn · TensorFlow · Keras · PyTorch | Spark · PySpark · Kafka · DuckDB · ETL/ELT | AWS (S3, Glue, Athena, QuickSight) · Tableau · Power BI |
Repo: 👉 aws-industrial-machine-failure-analysis
Stack: Python · PySpark · AWS S3 · Glue Crawler · Athena · QuickSight · DuckDB
- Ingested raw industrial sensor data (AI4I dataset) into S3 (raw + curated zones)
- Built a local PySpark ETL to transform CSV → Parquet with clean schema
- Used AWS Glue to discover schema & built a Data Catalog
- Queried curated Parquet via Athena and built a QuickSight dashboard with:
- Failure distributions
- Temperature vs failure behavior
- Torque and tool wear relationships
- Added a local DuckDB warehouse layer for fast analytics & BI export
Repo: 👉 geico-fraud-streaming-ml
Stack: Kafka · Spark Structured Streaming · Python · ML
- Simulated streaming insurance transactions
- Scored events with a trained fraud model in near real-time
- Logged predictions for monitoring & drift analysis
Repo: 👉 walbrydge-iot-predictive
Stack: Spark · ML · FastAPI · Docker · SHAP
- Predicts Remaining Useful Life (RUL) for equipment using time-series sensor data
- Serves predictions via FastAPI REST API
- Uses SHAP to generate ML explainability for engineers
- Packaged with Docker for portable deployment
Repo: 👉 AI-Banking-Assistant
Stack: LangChain · Vector DB · RAG · Python
- Built a Retrieval-Augmented Generation assistant for banking FAQ and policy Q&A
- Uses embeddings + vector search to ground answers in bank documents
Repo: 👉 model-drift-retraining
- Monitors feature & prediction drift
- Triggers retraining once drift crosses defined thresholds
- Adds a light MLOps layer on top of standard ML workflows
- 🧪 Fruit Maturity Detection Using MATLAB Image Processing
International Journal of Innovative Technology and Exploring Engineering (IJITEE) - 🌐 Comparative Analysis of IGP Protocols of an Enterprise Network
Paper Link
- Languages Used: Python, SQL
- Focus Areas: Data Structures, Algorithms, Dynamic Programming
- 📧 Email:
sampreethi4565@gmail.com
✨ Always open to collaborating on Data Engineering, ML, and Cloud Analytics projects.