Kshitish Mule
builds machines that see, hear,
and decide — offline.
Researcher in embedded AI, computer vision and edge robotics. B.Tech in Computer Science (AI & Data Science), 2022–2026 — currently preparing to pursue a Master's in Japan.
I'm an undergraduate researcher working at the intersection of embedded systems and applied AI — building autonomous systems that run entirely on-device, without depending on the cloud.
My final-year project, NURO, is a voice- and vision-controlled robotic system deployed on Raspberry Pi, combining real-time face detection, voice recognition and offline LLM inference into a single decision-making pipeline. It's registered with the Copyright Office, Government of India, and was presented at the 1st International Conference on AI Applications and Technological Innovations (ICAIATI-2025) in association with Taylor & Francis Group.
Alongside coursework, I've co-authored two published papers on machine learning and secure AI-to-device communication, and I've spent the last stretch of my degree learning Japanese to N4 level in preparation for graduate study in Japan, where I want to continue this work in a robotics or embedded-AI lab.
Interests & experience
What I study, and where I've applied it so far.
NURO — AI Robotics Research (Final-Year Project)
- Designed and built an autonomous robotic system integrating AI on Raspberry Pi with voice and vision control.
- Developed a real-time face-detection, voice-recognition and decision-making pipeline in Python and OpenCV.
- Deployed optimized TensorFlow Lite models on edge hardware; paper under review at Taylor & Francis.
- Copyright registered by the Government of India as "NURO: Neural Understanding for Robotics Operations" — Author & Owner #1 (Cert. No. LD-20260180327).
ML Algorithms Survey — Co-authored Research
- Co-authored a 17-page comprehensive review of machine learning algorithms, applications and future directions.
- Covered supervised, unsupervised and reinforcement learning, with analysis of real-world deployment challenges.
Selected builds
Hands-on systems, mostly aimed at running AI where the cloud can't reach.
NURO — Offline LLM Robotic System
2025 — Under developmentAn AI robotic system combining offline LLM inference, face recognition and voice commands on Raspberry Pi — full autonomous interaction with zero cloud dependency.
AI-Powered Assistive Robot
Raspberry PiAn assistive robotic platform using Raspberry Pi, built to support day-to-day tasks through sensor input and automated response.
Wearable Care
IoT / WearablesA wearable health-monitoring concept exploring sensor fusion for continuous, low-power tracking of vital signs.
ABD Prototype
Hardware / AutomationA prototype system exploring automated detection and response behaviour on embedded hardware.
Alpacca — Local LLM Inference
2024A C++ implementation of the Alpaca LLM for fully local inference without cloud APIs, with model optimization for limited hardware.
GPT Clone — AI Chatbot
2024A GPT-style chatbot with real-time conversation handling built on NLP and Flask, with a responsive web UI.
Papers & talks
-
Published · Review Article
Advancements in Machine Learning: A Comprehensive Review of Algorithms, Applications, and Future Directions
-
Published · 24 Dec 2025
Advanced Security Mechanisms for AIDL Communication in High-Risk Environments
-
Copyright registered · Filed 04 Nov 2025
NURO: An AI-Powered Voice and Vision Controlled Robotic System Using Raspberry Pi
-
Conference · 12–13 Dec 2025
1st International Conference on AI Applications and Technological Innovations (ICAIATI-2025)
-
Conference · 2025
Paper Presentation — STM Journals
Technical skills
Let's talk research.
Open to conversations about graduate programs, embedded AI labs, and robotics research in Japan.