Tashrique Ahmed

• Resolved CSS/JS cross-browser compatibility issues and optimized front-end performance while redesigning and developing the website using Next.js, TypeScript, React, MongoDB, Node, and Tailwind and integrated blog posting admin dashboard, JWT authentication, role-based access control, RESTful APIs, database schemas, and data caching strategies and parameterized queries to prevent SQL injection attacks
• Collaborated with the back-end team to optimize a weighted Hungarian algorithm for automatic driver assignment using Go, TypeScript and PostgreSQL, reducing double-assignment errors by 80% and manual interventions by 30%, saving 2 hours/week maintenance time
• Built JUnit tests in Java for finance transaction validation, covering 70% of workflows and edge cases with unit and integration tests, which reduced bug tickets by 15%, cut testing time by 30%, and improved release cycle efficiency by 20%
• Developed a subprocess for NetID collision resolution using C, SSMS and PostgreSQL, implementing automated tests following Test Driven Development (TDD) practices and containerizing with Docker, streamlining and reducing manual account management efforts by 30%

• Architected an SDLC-compliant automated update and data validation system using Python scripts for the Clerkships Database in Supabase (postgreSQL), reducing system downtime by 30% and saving approximately 8 hours of developer time per month
• Configured and deployed Grouper for Canvas admin management, integrating attestation features to enhance security and reduce unauthorized admin additions by 20%, utilizing Agile methodologies with Scrum and Jira, to streamline project execution and collaboration
• Migrated the team’s codebase to GitHub and established Agile version control practices, quickly learning and applying new methodologies to enable parallel feature development by other interns, improving project organization and reducing merge conflicts by 40%

• Implemented a Process Automation Tool using JavaScript, Java, and AWS S3 to automate a manual production process, boosting daily production by 570% (30 to 200 orders), collaborated with marketing team for market testing, projecting a 7% quarterly revenue growth
• Developed an OCR-based File Management System using Python, to automate folder tree creation and subject-based sorting by scanning receipts, reducing processing time by 15% (from 3 hours to 2.55 hours), increasing daily order capacity by 30
• Built and tested a convolutional recurrent neural network using TensorFlow and Keras to automate categorization of children’s artwork, achieving 92% accuracy, reducing sorting time by 60%, and boosting sales conversion rates by 15% through targeted recommendations

Soft Robot Inverse Kinematics for Virtual Reality - Prof. James Bern

• Restructured a Python-based algorithmic tool for kinematic pattern analysis of 3D Soft Robots using OpenCV and ROSPy, increasing simulation accuracy by 15%, providing visual insights that streamlined the next stages
• Led VR data collection and integration with a semantic-aware path planner, enhancing fluid turning navigation accuracy by 10%
• Modeled 3D components using SolidWorks, integrated layered elastomeric structures for tactile feedback, and pneumatic actuation channels for dynamic adaptability

• Developed the Fellowships page for Generation Conscious on their website, utilizing WordPress, CSS, and Javascript, managed coupons updates on the backend
• Directed the switch to environmentally safe detergents at Williams College, resulting in a notable 30% cut in water pollution and carbon emissions due to laundry
• Led a series of workshops and campaigns focused on sustainable laundry practices, effectively increasing student participation by 50% and fostering a community culture of eco-consciousness.

Fellowship's Page Web Dev

• Engineered a cost-effective automated turntable within $50 to replace manual handling of artifacts during Photogrammetry, and rapidly mastered Arduino, Fusion360, and EaglePCB within two days for this project
• Collaborated with archaeologists to digitize 10,000+ images of a 60-thousand-year-old cave bear tooth into a 0.1mm precise 3d model using photogrammetry, rendered and optimized clarity by 15% with Metashape and Photoshop, utilized ARkit and Unity for AR/VR interactivity, revealing two previously unknown morphological features about the fossil
• Led hands-on workshops at the Makerspace on 3D printing, laser etching, and CNC machines, teaching students to design, operate, and troubleshoot. Maintained and operated the 3D printers for various projects, ensuring optimal machine performance and print quality.

Paintbeat Art is the first multivendor art E-commerce Startup in Bangladesh which is a platform that celebrates and promotes various artistic expressions, including painting, sketching, crafts, calligraphy, digital arts, photography, and more. A group of 8 highschoolers started this project to promote local Bangladeshi art to international buyers.