Research Experience
Graduate Research Assistant
Sep 2024 – Sep 2025
Samsami's Lab · University of New Haven, CT
- Improved crack detection accuracy by 4% across multiple benchmarks by developing a custom Residual Vision Transformer (RvT) model in PyTorch for UAV-captured bridge imagery.
- Enhanced model interpretability by integrating LIME-based explainable AI methods to visualize prediction-relevant regions for infrastructure defect detection.
- Reduced false negatives by performing EDA on 40,000+ high-resolution images using OpenCV, Pandas, and NumPy.
- Contributed to 4 accepted papers: ASCE Journal, TRB 2026, CRC 2026, and 1 under submission.
Undergraduate Researcher
Sep 2022 – June 2023
University of Wolverhampton (Herald College) · Kathmandu, Nepal
- Developed and compared three CNN architectures (LeNet, ResNet50, VGG16) for breast cancer detection from histopathological images.
- Implemented data augmentation to address class imbalance, improving model generalization and robustness.
Industry Experience
Analytics Engineer
Jan 2026 – July 2026
Downtown Eveningx Soup Kitchen · New Haven, CT
- Conducted exploratory data analysis to identify trends and provide actionable insights supporting organizational decision-making.
- Designed and maintained data pipelines to collect, clean, and transform operational and donor data from multiplesources.
Graduate Assistant
Mar 2024 – Dec 2025
Graduate Admissions · University of New Haven, CT
- Developed Power BI dashboards with custom DAX measures and predictive analytics, reducing data interpretation time across departments.
- Managed communication workflows with hundreds of prospective students regarding admissions and program details.
Data Analyst Intern
Jan 2025 – May 2025
Downtown Evening Soup Kitchen · New Haven, CT
- Automated data workflows using Python, Zapier, and Excel macros, integrating 4 business systems and reducing reporting time.
- Implemented data validation protocols with Cognito Forms and MSSQL, improving accuracy across 30,000+ monthly records.
Associate Data Engineer
July 2023 – Dec 2023
Krispcall Pte. Ltd · Singapore
- Developed and optimized ETL pipelines, increasing data processing speeds for faster BI decisions.
- Analyzed user behavior using Mixpanel and Python scripts, identifying 20+ fraud instances and enhancing platform security.
Selected Projects
FinRAG: Retrieval-Augmented Generation for Financial Analysis
- Multi-source RAG system combining 289,642 documents from Reddit, historical stock data, and financial literature.
- Benchmarked Mistral-7B, Phi-3-Mini, and Llama-3.1-8B on domain-specific financial query generation.
Data Engineering Pipeline with AWS & Apache Spark
- Scalable ETL pipeline using AWS Lambda and Apache Spark for distributed analysis of large-scale YouTube dataset.