Headline

GSoC’25 @ Human AI | Ex-AI Intern @ Krop.ai & Kalppo | Ex-Intern @ Vanora Robots | 3x National Hackathon Winner | Multi-Agent Systems | Research-Driven Builder

About

I’m a Mechanical Engineering student at NITK with a minor in Machine Learning, interested in the kind of work that sits at the edge of design, code, and systems thinking. I’ve built everything from multi-agent NLP pipelines for legal document analysis to real-time fake news detectors, and even dabbled in OCR-based document classifiers. A lot of that lives on my GitHub, where I keep experimenting with transformers, FastAPI, LangChain, spaCy, and other cool stuff in the ML/NLP space. On the other side of things, I’ve worked on mechanical design projects like building a gas pump prototype for a client using SolidWorks — including detailed part design, assembly, and simulation. I like projects that have both physical and computational elements, and I’m always looking to tie the two worlds together. Outside of tech, I’m into sports (basketball mostly), and generally just like building things that are either useful, clever, or fun.

Location

Mangaluru

AI stack

anthropic apis

claude code

cursor

gemini

hugging face

claudecode

Work experience

ML Engineer at Fermions, May 2026 - Sep 2026 · 4 mo, Remote

Artificial Intelligence Intern at Krop AI, Nov 2025 - Feb 2026 · 4 mos, Udupi, Architected scalable telemetry data pipelines and an Over-the-Air (OTA) update system for autonomous farming containers. Built the core infrastructure bridging hardware and software to enable real-time remote management and data flow.

Research Fellow at Google Summer of Code, May 2025 - Oct 2025 · 6 mos, Remote, • Designed full NLP pipeline for MIC detection at Google Summer of Code • Extracted conflict actions & fatality patterns to enhance data analysis • Boosted accuracy using heuristic + ML fusion techniques for improved results

AI Intern at Kalppo, Apr 2025 - Jul 2025 · 4 mos

Design and Modelling Intern at Vanora Robots, Aug 2024 - Apr 2025 · 9 mos, Mangaluru, Karnataka, India, Designed and modeled a client-specific gas pump prototype using SolidWorks, focusing on functionality, manufacturability, and design aesthetics. Collaborated with ML engineers to integrate CAD outputs into a defect detection simulator, contributing to a 40% reduction in manual QC effort. exposure to client-driven product development

Education

Minor, Machine Learning, National Institute of Technology Karnataka

Bachelor's degree, Mechanical Engineering, National Institute of Technology Karnataka

Public links

github: https://github.com/Joshuathomas18