karthik

Karthik Alapati

karthikgatiganti@gmail.com Hyderabad, India CV LinkedIn GitHub

About

I am a senior student majoring in Computer Science (AI) at Jawaharlal Nehru Technological University, Kakinada (JNTUK). I'm currently working as a Machine Learning Research Intern at IIT Kharagpur. My main research focus includes LLMs, Multimodal LLM Agents and Secure ML.

My work experiences, though varied in setting, are united by a common thread: a drive to apply my skills in machine learning and software development to solve real-world problems. My most recent role, as a Machine Learning Intern at FoodCLUB (Feb 2024-Present), focuses on integrating AI-powered features into their Android application to enhance user experience. I'm currently exploring various ML models to personalize recommendations and streamline the ordering process. Concurrently, I am also a Machine Learning Intern at LILLUP (Feb 2024-Present), where I'm investigating innovative ways to leverage machine learning within their Android client, aiming to improve user engagement and satisfaction.

Prior to these roles, I worked as a Founding Member of Arlarse (Mar 2024-Present), a free AI learning platform. This experience allowed me to contribute to a project with a strong social mission – democratizing access to education. My responsibilities included platform development and content creation, focusing on making complex AI concepts accessible to a wider audience. During the summer of 2023, I interned as a Machine Learning Intern at Indian Servers, where I developed a Hair Health Care LLM by fine-tuning GPT-2 on a specialized medical dataset. This involved data preprocessing, model training, and integrating the model with a Telegram bot for user interaction. Earlier that year, I also contributed to open-source development as a Linux Kernel Developer Mentee at The Linux Foundation (Jun 2022 - Aug 2022), debugging the kernel and submitting patches to improve Coccinelle.

Research Interests

Publications

Publication Image

GIC: Secure and Efficient classification of Mobile Gallery Images

Karthik Alapati, 2024

Projects

Publication Image

NAC: Neural Aided Compression

A project backing the SubSampler app. Subsampler is a lightweight app that helps you compress images to the required size. It is a great tool for anyone who needs to compress images for irritating websites with unreasonable size requirements, social media, or other purposes. It is also a good way to save space on your phone's storage.

Publication Image

Flasher: Linux ISO image flash tool for Android

An app for flashing Linux ISO images(or any) to pen drive with 1 click. It has over 100K downloads on Google Play Store, enabling users to flash Linux ISOs without using desktops. It leverages raw USB Linux interface to flash 5GB+ ISO images to the USB drive using SCSI protocol without root.

Publication Image

Magic Files: libmagic based file manager

The only file manager that detects file-type based on its magic number using libmagic. It is the only File manager for Android that detects 1000+ file types and has support for more than 10 Archive and Compression formats by integrating libarchive.

Publication Image

Helper: A 3 in 1 assist tool for developers

An App killer, App activities explorer, App apk extractor. It packs App activities launcher, Memory monitor and App Killer into 1 App to help developers during development. It leverages accessibility-service to simulate a user click and 3+ different heuristics to put the app in stopped state.

Free Software

I am a Debian Developer from 2020. I built and currently maintaining 7 packages across 3 teams including Deep Learning and Python teams for Debian.