Turning messy data into actionable business insights through AI, machine learning, and analytics.
I build things at the intersection of Analytics, LLMs, and Agentic AI.
Before grad school, I spent 5 years in industry as a Senior Data Scientist at eClerx and a Data Scientist at UST, working on ML systems, analytics pipelines, NLP search, forecasting, predictive maintenance, and business-impact data science.
I recently completed my MS in Computing — Artificial Intelligence Track at the University of Utah. I also interned at BMW Group while completing the course.
🟢 Currently looking for full-time Data Scientist / ML Engineer / AI Engineer roles in the US.
- LLM & RAG systems — semantic search, retrieval pipelines, summarization, recommendations, and context-aware applications
- AI agents — multi-agent workflows, tool-using systems, and reasoning-based recommendation pipelines
- Computer vision models — object detection, segmentation, medical imaging, and damage detection
- Applied ML systems — forecasting, predictive modeling, anomaly detection, customer segmentation, and business analytics
- Data pipelines — SQL, Databricks, Spark, ETL workflows, dashboards, automation, and reporting systems
| 5+ years Data Science & AI experience |
$100M–$300M Monthly originations supported through BMW loan optimization pipeline |
600+ Credit attributes processed |
$50M Missed business opportunity identified at eClerx |
| 2% MAPE Order conversion forecasting |
98% Medical-bed ML pipeline accuracy |
200% Analysis efficiency improvement |
0.64 mIoU Coronary vessel segmentation |
| Area | Tools |
|---|---|
| ML / AI | LLMs, RAG, Agentic AI, NLP, Computer Vision, Deep Learning, Statistical Modeling, Predictive Modeling |
| Frameworks | PyTorch, LangChain, Scikit-learn, OpenCV, Hugging Face, SpaCy, NLTK, Gradio, Flask |
| Data & Infra | Databricks, Spark, SQL, BigQuery, MySQL, Oracle, ChromaDB, Docker, CI/CD, Power BI |
| Languages | Python, C++, R, SQL |
|
Multi-agent GenAI system that recommends music from images using a RAG Pipeline with multi-agent system. It uses vision-language reasoning to understand visual context and generate track recommendations by relying rules from the Knowledge Base. Stack: LangChain · RAG · Agents · Vector DB · Multimodal GenAI 
|
Fine-tuned Faster R-CNN to detect guardrail damage from dashcam images. Applied Focal Loss for class imbalance and evaluated model performance using GIoU and mAP. Also suggested a method to reduce false positives Stack: PyTorch · Faster R-CNN · Computer Vision · Fine-tuning
|
|
RAG-based book recommender that suggests books from user-provided topics, categories, and sentiment preferences using embeddings and a Gradio interface. Stack: RAG · LLMs · LangChain · Hugging Face · Gradio · Python 
|
U-Net-based semantic segmentation model for coronary vessel trees from X-ray coronary angiography images, achieving up to 0.64 mIoU. Stack: PyTorch · U-Net · Medical Imaging · Segmentation · CV
|
BMW Group — Data Science Intern
Built scalable data pipelines with advanced sampling and cleaning for 600+ credit attributes, enabling a loan optimization model tied to $100M–$300M in monthly originations.
eClerx — Senior Data Scientist
Worked on large-scale analytics, forecasting, LLM prototypes, Databricks and Power BI dashboards, ETL automation, customer segmentation, and business-impact analysis.
Highlights:
- Identified $50M in missed business opportunities from delayed product listing data
- Forecasted web and app conversion rates using ETS, ARIMA, and Prophet with 2% MAPE
- Built LLM-based prototypes for retail review summarization and emotion detection
- Architected scalable ETL workflows and automated reporting, improving analysis efficiency by up to 200%
UST / Abzooba — Data Scientist
Worked on NLP search, predictive maintenance, customer analytics, web journey analytics, and ML pipeline optimization.
Highlights:
- Built BERT-based medical search over clinical text
- Built predictive maintenance models using real-time sensor data from silicon photonics machines
- Engineered a Python ML pipeline for patient position prediction with 98% accuracy
- Optimized an ML pipeline into C++ for System-on-Chip execution
- Building stronger RAG systems and practical AI agent workflows
- Improving computer vision model performance on real-world data
- Turning ML prototypes into clean, usable applications
- Looking for full-time Data Scientist / ML Engineer / AI Engineer roles in the US
I like building systems that are useful, not just technically interesting.
Outside of work, I’m into drums, percussion, running, cycling, and occasionally going deep into anime rabbit holes.
Happy to connect — whether it’s about a role, a collaboration, or just talking about applied AI.
