Share Profile
SYED MUHAMMAD ATHER HASHMI
Aspiration Statement
My aspiration is to become a leading Data Scientist and Deep Learning expert, contributing significantly to the realm of Large Language Models and Software Engineering.
Core Skills
- C++, Database Management (SQL, MongoDB), Large Language Model training and Finetuning, Python
Academic Awards / Achievements
- Excellence Scholarship throughout 4 years of study
Experience
Leadership / Meta-curricular
- General Secretary, Pride Press Journalism and Publication Club Editorial Head, Pride Press Journalism and Publication Club Content Lead, Google Development Student Club, HU Chapter Digital Ambassador Student Life Ambassador
Internship / Volunteer Work
- AI Datasphere Intern, Blutech Consultings (June 2024 July 2024) Teaching Assistant for CS101 Algorithmic Problem Solving, Habib University (August 2023 December 2023) Information Security Intern, Habib Bank Limited (June 2023 ��� July 2023) Section Leader and CS Teacher, Stanford University (April 2023 ��� June 2023) Teaching Assistant MATH205 Linear Algebra, Habib University
Publications / Creative Projects
- Finetuning Llama 3.2 1B model for Habib University FAQs
Final Year Project
Project Title
Drone Image Stitching
Description
The project focuses on drone image stitching, aiming to create high-resolution, geo-referenced panoramic images from aerial photographs. Utilizing advanced feature detection and matching techniques such as SIFT, SURF, ORB, SuperGlue, LoFTR, and Dedode, the project seeks to optimize efficiency and reduce latency in large-scale aerial mapping. The stitching process involves image metadata extraction, feature matching, homography estimation, and seamless image blending to ensure accurate alignment and minimal distortions. This research benefits applications in agriculture, environmental monitoring, and urban planning, where large-area mapping is essential. though challenges arise with larger images with fewer keypoint matches. By refining pairwise stitching strategies and optimizing computational resources, this project aims to enhance real-time drone imaging, making large-scale mapping more precise, efficient, and scalable for diverse applications.