About Me
Hello! My name is Ramana and I enjoy building things and creating experiences that make an impact. I consistently strive to improve, to think differently, and to contribute to something bigger than myself. I'm currently pursuing a Bachelor of Technology in Computer Science & Engineering at VIT, Vellore, graduating in 2026.
Fast-forward to today, and I've had the privilege of working at Samsung Research as a Research Intern, where I built internal solutions to fast track debugging. I'm also an AI/ML Intern at GloballyGI, where i'm working on computer vision and model inference optimization. Previous summer, I interned at NIT-Trichy, Before that also as Software Engineer & Research Analyst for Team Levitate Hyperloop, winning at the Global Hyperloop Competition 2025 at IIT-Madras. I was also a part of ACM-VIT one of the most prestigious technical chapters at VIT Vellore.
I also won a National Hackathon at IIT-Gandhinagar (competing against 3,000+ teams) for building NIC-Explorer, a semantic search engine achieving 90% accuracy with support for 12+ Indic languages.
I have experience working with a diverse range of technologies and frameworks, including full-stack mobile application and web development, building custom software packages, developing machine learning and deep learning applications. I am currently expanding my expertise by exploring blockchain technologies.
Where I've Worked
Research Intern @ Samsung Research
Sep 2025 - Present
- Reduced telecom debugging time by 99%, from days to seconds by building an automated Call Flow Visualizer that parses multi-source network logs and renders end-to-end sequence diagrams
- Architected a deterministic parsing pipeline achieving 100% reproducibility and zero false positives compared to ML solutions, eliminating costly production debug failures across multiple testing teams
- Designed a scalable backend for an agentic LLM analysis system to handle large log files through intelligent chunking and selective routing, avoiding context window limitations while maintaining deep analytical accuracy
Some Things I’ve Built
Featured Project
Mac-Watcher
Automated macOS security monitoring CLI capturing webcam, screenshots, location, and network data on system wake. Features asynchronous processing, email alerts via Resend API, and offline queuing. Production-ready with Homebrew install.
- Shell
- Objective-C
- Ruby
- macOS
![Mac-Watcher]()
Featured Project
NIC-Explorer
Semantic search engine for NIC codes using Graph RAG with Neo4j, achieving 90% accuracy. Supports 12+ Indic languages with speech-to-text. Won National Hackathon at IIT-Gandhinagar (3,000+ teams) for highest accuracy and best UI.
- Next.js
- Neo4j
- FastAPI
- Groq
- RAG
![NIC-Explorer]()
Featured Project
GitHub-Painter
Open source tool enabling users to create custom patterns on GitHub contribution graphs through commit history manipulation. Contributed multiple PRs adding new features and improving script functionality.
- JavaScript
- Shell
![GitHub-Painter]()
Other Noteworthy Projects
Cobe
Contributed to popular 5kB WebGL globe library used in 9,000+ projects. Added 7 new API functions for improved physics manipulation and enhanced simulation playground for real-time option testing.
LoginWatcher
Monitors macOS login attempts (TouchID/password) and triggers custom scripts on success/failure. Runs in background at startup, passes contextual data via environment variables, and optimizes monitoring for locked screens.
WiFiWatcher
Monitors Wi-Fi network changes and executes custom scripts on connection/disconnection events. Features real-time SSID monitoring, pattern matching, context-aware environment variables, and advanced condition combinations for complex automation.