Hi đź‘‹ My name is Saksham Mrig (sksum), I am a 3rd year Computer Science and Applied Mathematics undergraduate at IIIT Delhi, I am an enthusiastic young developer who can google like a pro.
Publications
Leveraging Intra and Inter Modality Relationship for Fake News Detection
Proposed a novel model architecture that leverages the fine-grained salient image and text features that outperformed existing methodologies by an average of 3.05% [accuracy] and 4.53% [F1-score] on two publicly available datasets
Developed a neural network based recommendation system that recommends the best on-site to cloud migration strategy tailored for each customer with a MAP@[k=3] score of 0.70 Synthesized the data into 20+ executive-ready insights, Lead the dialogue with stakeholders and senior team members on how to generate firm-wide impact. Google Summer of Code (with SugarLabs, Boston)
Resolved 72 issues in MusicBlocks along with major enhancements such as: Integrated USB MIDI input, PitchAnalyser (Audio to Pitch converter), and also improved the in-app search results.
Redesigned the official IIITD library portal using SAPPER and headless Drupal, decreasing the page load time by 32%. Devised an easy-to-use system for website managers, enabling them to create responsive and interactive web pages with plain markdown and yaml files
A concurrent, lightweight, multi-threaded webserver made in C with integrated SvelteKit and libtorch to deploy modern web interfaces for machine learning models, either trained in or exported to C++
Built a chrome extension and developed a self-trained mood detection engine using TensorFlow and Keras achieving 76% accuracy. Tracked users’ browsing session using metadata and webpage keyword extraction algortihms like RAKE, increasing users’ productivity.