Devansh Lodha

Ko Tapu, Phang Nga Bay

(from "The Man with the Golden Gun")

Hello! I am an undergraduate student at IIT Gandhinagar (Class of 2027) pursuing B.Tech. in Computer Science and Engineering.

My fascination with Probability and Statistics, sparkled by Professor Shanmuga’s class during my second semester led me to delve deeper into Machine Learning. I am interested in Computer Vision. I’ve previously worked at the Sustainability Lab under Professor Nipun Batra on Vision Language Models for Object Detection in Satellite Imagery.

Similarly, my fascination with Systems and Networks ignited by Professor Sameer G. Kulkarni’s Computer Networks course as well as Professor Abhishek Bichhawat’s Operating Systems course led me to work on AI-Native Network Diagnostics at DoT 5G Use Case Lab as well as contributing to Rust Embedded’s OS Tutorials on the Raspberry Pi!

Outside of academics, I love relaxing through reading, running and gaming.

When I need a break, I love getting lost in sci-fi movies, especially the work of Christopher Nolan. Music is also a constant companion; you can find me listening to a variety of genres. Here are some of my absolute favorites:

news

Jan 26, 2026	Made it to Semester-5’s Dean’s List! (Letter of Appreciation)
Jan 07, 2026	I presented my first paper, “MCP-Diag: A Deterministic, Protocol-Driven Architecture for AI-Native Network Diagnostics,” at IEEE COMSNETS 2026 and received the Best Graduate Forum Paper (Runner-Up) award!
Aug 07, 2025	Selected for Google India’s 2026 Summer Internship!
Jun 04, 2025	I delivered a tutorial on Vision Language Models for the AI for Social Good ACM India Summer School. I enjoyed the session with the 60+ participants. For those interested, it available on YouTube and on my Blog
Mar 20, 2025	Selected for the Summer Research Internship Program (SRIP) 2025 at IIT Gandhinagar.
Feb 25, 2025	Recognized by the Head of Department (HOD), CSE, IIT Gandhinagar, for being among the top 20% of students in the Computer Science and Engineering branch. (Letter of Acknowledgement)
Sep 04, 2024	Joining TEDxIITGandhinagar as Licensee, Organizer and Design Core!
Aug 15, 2024	Made it to Semester-2’s Dean’s List! (Letter of Appreciation)
Jan 26, 2024	Made it to Semester-1’s Dean’s List! (Letter of Appreciation)

selected publications

MCP-Diag: A Deterministic, Protocol-Driven Architecture for AI-Native Network Diagnostics

Devansh Lodha, Mohit Panchal, and Sameer G. Kulkarni

2026

Best Graduate Forum Paper (Runner-Up) Abs Bib PDF

This paper received the Best Graduate Forum Paper (Runner-Up) award at IEEE COMSNETS 2026.

The integration of Large Language Models (LLMs) into network operations (AIOps) is hindered by two fundamental challenges: the stochastic grounding problem, where LLMs struggle to reliably parse unstructured, vendor-specific CLI output, and the security gap of granting autonomous agents shell access. This paper introduces MCP-Diag, a hybrid neuro-symbolic architecture built upon the Model Context Protocol (MCP). We propose a deterministic translation layer that converts raw stdout from canonical utilities (dig, ping, traceroute) into rigorous JSON schemas before AI ingestion. We further introduce a mandatory "Elicitation Loop" that enforces Human-in-the-Loop (HITL) authorization at the protocol level. Our preliminary evaluation demonstrates that MCP-Diag achieving 100% entity extraction accuracy with less than 0.9% execution latency overhead and 3.7x increase in context token usage.
@misc{lodha2026mcpdiag, title = {MCP-Diag: A Deterministic, Protocol-Driven Architecture for AI-Native Network Diagnostics}, author = {Lodha, Devansh and Panchal, Mohit and Kulkarni, Sameer G.}, year = {2026}, archiveprefix = {arXiv}, primaryclass = {cs.NI}, url = {https://arxiv.org/abs/2601.22633}, }