AIMLInnovations

Projects & Experience

Explore selected analytical projects and professional experience that highlight research, software engineering, and data-driven impact.

Experience

Graduate Research Assistant

Walker Institute @ Western Michigan University - Kalamazoo, USA

August 2023 - Present

Designed and deployed data collection pipelines and survey instruments to measure student outcomes across STEM education programs, enabling quantitative evaluation of program effectiveness.
Built and validated psychometric models across multiple domains (math proficiency, engineering design, STEM awareness), supporting structured and scalable outcome assessment.
Performed statistical modeling and regression analysis in R/Python to identify key predictors of student performance and quantify program impact.
Applied advanced analytical methods, including functional data analysis and network analysis, to uncover relationships in longitudinal participant data.
Developed data-driven evaluation frameworks and baseline metrics to support continuous monitoring, reporting, and iterative program improvement.
Translated complex analytical findings into actionable insights and recommendations for stakeholders, influencing curriculum design and program strategy.

Software Engineer

Deductive Clouds - Bangalore, India

May 2021 - Jul 2022

Designed and implemented a 3D spatial optimization algorithm for logistics operations, improving packing efficiency and resource utilization by up to 37%.
Developed and maintained backend services and APIs using Node.js/Express, enabling secure data access, system integration, and scalable application performance.
Led development of an end-to-end logistics and inventory management system, integrating 3D bin packing, ERP workflows, and document processing features.
Refactored and enhanced mobile applications using React Native, including Google Maps integration for driver workflow optimization.
Implemented frontend state management (Redux) and improved UI responsiveness for inventory and delivery tracking systems.
Coordinated and mentored a small engineering team, improving development efficiency and reducing overall delivery timelines.

Projects

Data science and analytics projects leveraging reproducible workflows, statistical modeling, and performance evaluation.

Analyzed large-scale educational assessment data to evaluate the impact of special education service intensity on standardized student performance outcomes.
Performed statistical modeling (ANOVA, t-tests, regression) to identify performance disparities across grade levels, demographics, and student groups.
Engineered features and segmented cohorts to uncover statistically significant trends and quantify outcome differences.
Developed reproducible data analysis pipelines in R/Python to improve transparency, consistency, and repeatability of insights.
Translated analytical findings into actionable recommendations to support data-driven decision-making in educational program design.

Data science and analytics projects leveraging reproducible workflows, statistical modeling, and performance evaluation.

Built a data analysis pipeline to explore temporal and categorical trends in YouTube video performance across regions and content types.
Performed exploratory data analysis and feature engineering to identify key drivers of views, engagement, and content virality.
Applied statistical analysis and data visualization techniques to uncover patterns in audience behavior and content popularity.
Optimized data processing workflows for large datasets, improving runtime efficiency and scalability.
Generated insights to support trend forecasting and content strategy decisions.

Data science and analytics projects leveraging reproducible workflows, statistical modeling, and performance evaluation.

Built an end-to-end 3D reconstruction pipeline using Structure-from-Motion (SfM) to generate 3D scenes from large-scale 2D image datasets.
Implemented feature extraction and matching using SuperPoint and SuperGlue to improve robustness of correspondence detection across varying viewpoints and lighting conditions.
Integrated a ResNeXt50 deep neural network for geometric verification and camera pose correction, improving reconstruction accuracy.
Optimized pipeline performance using half-precision computation and caching techniques, reducing memory usage and improving scalability.
Automated reconstruction workflows using COLMAP, enabling efficient generation of high-quality 3D environments based on match and pair selection thresholds.