Skip to content
View DevasishK's full-sized avatar
💻
Open to SDE, AI/ML & Cybersecurity Opportunities
💻
Open to SDE, AI/ML & Cybersecurity Opportunities

Block or report DevasishK

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
DevasishK/README.md

GitHub Banner

Hi 👋, I'm Devasish Viswanadh Kolla

AI Researcher • Cybersecurity Enthusiast • Full-Stack Developer

PortfolioLinkedInEmail


🚀 About Me

🎓 Final-Year B.Tech Computer Science Student @ SRM University AP

🔬 IEEE Published Author

📜 Patent Applicant in Trustworthy AI

☁️ AWS Certified Cloud Practitioner

🛡️ Cybersecurity Intern @ APTS & APCSOC (Govt. of Andhra Pradesh)

🤖 Passionate about AI, NLP, Explainable AI, Cybersecurity, Cloud Computing, and FinTech

🌱 Currently working on:

  • Loan Default Risk Analyzer using Explainable AI
  • AI Hallucination Detection Research
  • Secure Backend Systems with Spring Boot
  • Web Application Security Assessments

🏆 Achievements

  • 📄 IEEE Published Author
  • 📜 Patent Application Filed
  • ☁️ AWS Certified Cloud Practitioner
  • 🎓 Merit Scholarship (75% Tuition Waiver)
  • 🥉 Bronze Medalist – Karate Kata
  • 🛡️ Cybersecurity Intern – Government of Andhra Pradesh

📚 Research & Publications

📄 The Perils of Naive Truncation: A Context Ablation Study for Dialogue Summarization on DialogSum

  • Published at IEEE AISP 2025
  • Evaluated dialogue truncation strategies using transformer models
  • Analyzed effects of context ablation using ROUGE and BERTScore

📌 Patent Work

System and Method for Detecting Hallucinated Information in AI-Generated Summaries Using Multi-Stage Verification Architecture

  • NLI-based verification
  • Composite Faithfulness Risk (CFR) Score
  • Multi-stage hallucination detection
  • Sentence-level fact validation

🌐 Connect With Me


💻 Tech Stack

Java Python C++ SQL

PyTorch Transformers Scikit-Learn XGBoost SHAP LIME

Spring Boot FastAPI PostgreSQL MySQL

React TypeScript Tailwind CSS Vite

OWASP Burp Suite OWASP ZAP Wireshark

AWS Docker Git GitHub Actions Linux


📊 GitHub Statistics


🔥 GitHub Streak


🚀 What I've Been Building

🤖 Faithful Text Summarization

Explored the reliability of AI-generated summaries using BART, ROUGE, BERTScore, and NLI-based faithfulness evaluation. Focused on identifying hallucinations and improving trust in Generative AI systems.

📄 Dialogue Summarization Research

Author of the IEEE publication "The Perils of Naive Truncation: A Context Ablation Study for Dialogue Summarization on DialogSum". Investigated how context loss impacts summary quality in transformer-based dialogue summarization.

🔐 Secure File Locker

Engineered a cryptography-driven file protection system in C++ using AES encryption, SHA hashing, and secure access controls to ensure confidentiality and integrity.

🔑 JWT Authentication Backend

Designed a secure Spring Boot backend implementing JWT-based authentication, authorization, and role-based access control for modern web applications.

💰 Loan Default Risk Analyzer

Building an Explainable AI framework for NBFCs using ensemble learning, SHAP, and LIME to make credit risk predictions transparent and actionable.

🌐 Personal Portfolio

Crafted a modern, interactive portfolio showcasing research, projects, certifications, and technical expertise with React, TypeScript, Tailwind CSS, and Vite.


🎯 Areas I Love Exploring

🧠 Artificial Intelligence & Machine Learning

🔍 Explainable AI (XAI)

📚 Natural Language Processing

🛡️ Cybersecurity & Secure Software Development

☁️ Cloud Computing & Backend Systems

💹 FinTech & Data-Driven Decision Making


💡 My Mission

To build trustworthy AI systems, secure software, and scalable technology solutions that solve meaningful real-world problems while maintaining transparency, reliability, and user trust.

Pinned Loading

  1. faithful-text-summarization faithful-text-summarization Public

    Comparative evaluation of BART-based abstractive summarization on CNN/DailyMail and XSum datasets using ROUGE, BERTScore, and NLI-based faithfulness assessment.

    Jupyter Notebook

  2. JWT-Authentication-SpringBoot JWT-Authentication-SpringBoot Public

    JWT-based authentication and authorization system built with Spring Boot 3, Spring Security, and JSON Web Tokens for secure REST APIs.

    Java

  3. Secure-File-locker-in-CPP Secure-File-locker-in-CPP Public

    Secure File Locker is a command-line tool written in modern C++17 that lets you safely encrypt, decrypt, verify, and manage files with password-based protection. It uses industry-standard cryptogra…

    C++

  4. Shop_forage Shop_forage Public

    A full-stack e-commerce management system built with React and Spring Boot.

    JavaScript

  5. dialogue-summarization-context-ablation dialogue-summarization-context-ablation Public

    Context ablation study for dialogue summarization on DialogSum, analyzing the impact of naive truncation using ROUGE and BERTScore.

    Jupyter Notebook

  6. Personal_Portfolio Personal_Portfolio Public

    Interactive personal portfolio featuring AI research, cybersecurity projects, full-stack development, and a dual-mode user experience.

    JavaScript