At Mercor, I engage with international clients to create innovative data-driven solutions that enhance decision-making and operational efficiency. My role encompasses the entire data science lifecycle, from exploratory analysis to deployment, ensuring that our clients benefit from automation and improved processes. This experience has sharpened my skills in data engineering and modeling, allowing me to contribute effectively to diverse projects.

– Designed and deployed an AI-driven summarization pipeline using LLM OpenAI GPT-4o to transform raw web accessibility scan data

into structured insights, reducing manual review overhead for non-technical stakeholders.

– Engineered a backend system in Node.js to dynamically extract report content from the DOM and trigger summarization APIs,

integrating real-time inference into production data workflows.

– Developed a data-centric scoring system using Python and OpenAI’s CLIP model to assess image alt-text accessibility by combining

visual embeddings with contextual text data, aligned with WCAG 1.1.1 standards.

– Validated pipelines across public datasets from Harvard Accessibility and The Verge, and exported benchmarking results to CSV to

support accessibility analytics and ML model refinement.

• Spearheading the development of a modular and efficient pipeline for Image Analysis Performance Curves (IAPC), optimizing detection and classification in high-resolution satellite imagery.

• Implementing YOLO and TensorFlow models to enhance object detection capabilities, specifically for the xView dataset, and improve computational efficiency.

• Collaborating with FAST Labs to conduct research on knee detection algorithms in performance curves, contributing to potential publications.

• Conducting detailed experiments with synthetic aperture radar (SAR) and other image data types to validate the pipeline's adaptability and accuracy.

• Presenting findings through comparative analysis and working on scalable solutions for future applications in satellite image processing.

– Built and deployed a scalable ETL pipeline using Python and SQL to consolidate data from Shopify, Instagram, ERP, and internal

systems, streamlining reporting across departments and reducing aggregation time by 40%.

– Developed Power BI dashboards integrating sales, shipping, and ad spend KPIs, enabling operational managers to monitor marketing

ROI and improve efficiency by 20% through quicker actionable insights.

– Performed time-series forecasting on seasonal product categories using Python, achieving R2 = 0.79 and enabling inventory planners to

optimize stock and boost forecast accuracy by 20%.

– Collaborated with marketing to build a monthly campaign analytics pack that summarized reach, click-through rates, and attribution

paths, improving reporting clarity and executive visibility.

– Analyzed sales performance across 12 product categories using Python and SQL, uncovering revenue-impacting insights that led to a

10% increase in quarterly revenue through informed strategy revisions and reporting.

– Developed and automated executive dashboards using Tableau and Power BI to track KPIs and enable stakeholder access to real-time

trends, reducing manual reporting time by 30% and improving decision velocity.

– Implemented A/B testing to compare conversion flows, analyzed behavioral patterns using SQL and Python, and delivered data-backed

UX recommendations that increased user retention by 15% quarter-over-quarter.

– Created sales funnel analysis reports to identify customer drop-off points across checkout stages, contributing to a 12% improvement in

conversion rate through targeted UX fixes.