
Shaeelhashmi
Web developer
Kompetenzen

Meine Dienstleistungen


Portfolio
Arbeitserfahrung
Machine Learning Developer
LinkedIn • Vollzeit
Aug 2025 - Oct 2025 • 2 mos
During my time at the company, I focused on building and optimizing end-to-end deep learning pipelines. I architected and implemented a fine-tuning pipeline using PyTorch to adapt a state-of-the-art automatic speech recognition model for Hindi speech-to-text, focusing heavily on enhancing transcription accuracy and processing efficiency. In addition to speech processing, I designed and developed multiple computer vision applications. This included engineering a Convolutional Neural Network (CNN) based image character classifier optimized for structural pattern recognition and automated data extraction. Beyond core model development, I bridged the gap between machine learning and scalable system design. I conducted comprehensive technical research and delivered key architectural insights on how to implement secure, isolated, and highly performant multitenant architectures to support scaling infrastructures.
Automation Developer
Ninjas Code • Teilzeit
Jan 2025 - Mar 2025 • 2 mos
During my time at the company, I focused on improving repository visibility and maintaining development workflow hygiene across a portfolio of forked repositories. I engineered a systematic approach to boosting SEO performance by strategically creating and managing issues across forked repositories, opening approximately 40 issues daily to drive engagement signals to the original codebases. In parallel, I owned the dependency management lifecycle for these repositories, reviewing and merging around 40 Dependabot pull requests per day to keep dependencies current and minimize security vulnerabilities. To scale this workflow sustainably, I designed and built an automation script that streamlined the issue-creation process end-to-end, reducing manual effort by approximately 30% and freeing up time for higher-value tasks. Beyond day-to-day execution, I established a data-driven feedback loop by tracking and maintaining repository traffic metrics using Google Sheets, enabling consistent visibility into performance trends over time.