Ashwith Poojary

About Me

“

A robot without AI is like a body without a brain. A brain without a body lacks purpose.

- Ken Goldberg

Hi, I'm Ashwith Poojary - a robotics software engineer driven by the art of giving machines a mind and a mission. My world? Building systems that move, learn, and adapt with intelligence baked into every wire.

From robotic arms to deep learning swarms, I bridge the gap between hardware and intelligence. My true passion? Ideas brought to life in silicon and steel.

When I'm not coding, you'll find me writing articles or tinkering with servo motors. Welcome to my world where robots come alive.

Resume

Education

Master's in Robotics and Autonomous Systems (concentration: Artificial Intelligence)

Arizona State University | 2024-2026

Bachelor's in Computer Science

University of Mumbai | 2018-2022

Publications

Published a project on Indian sign language translation in International Research Journal of Engineering and Technology (IRJET)

Published a survey paper on Machine Learning approach to predict the trends of COVID-19

Working on publishing research findings on Domain generalisation and Medical Image Analysis through knowledge imputation and RAG framework in Neural Information Processing Systems (NeurIPS) and Empirical Methods in Natural Language Processing (EMNLP)

Skills

Programming Languages

Python Java TypeScript MATLAB

Frameworks/Technologies

Jetpack Compose React Native Flutter LangChain ROS2

Tools

Git PyTorch OpenCV Arduino

Areas of Expertise

Machine Learning Robot Learning Control Systems Android Development

Download Resume

Experience

Software Development Engineer

MyJio | Reliance Jio Platforms Limited | 2022-2024

I boosted the MyJio app's performance by 30% with Jetpack Compose, focusing on stable data and cool UIs

Led a React Native team of 10 to build a fintech app from scratch, creating in-house libraries and automating tasks to save both time and money

I mentored some associate engineers and held sessions for the team on Android optimization, Jetpack Compose animations, and Kotlin database tricks

Machine Learning Intern

AU Small Finance Bank | 2021

Built a machine learning model to predict loan eligibility with 90%+ precision, making marketing more efficient and cutting down resource waste

Used Polars to clean up and analyse over 500k customer records, which improved data consistency and bumped up model accuracy

Put together a report on how the model helps with loan approvals and marketing, making it easier for the team to make data-driven decisions

Mobile App Intern

JK Tyres | Applore Technologies | 2020

Worked on a hybrid Flutter app from scratch for both Android and iOS, using BLoC for clean state management and better performance

Collaborated with backend folks on security and architecture discussions — even got appreciated for helping tighten things up

Designed screens super fast, pitched in during crunch time, and helped ship the app on time under tight deadlines

Blogs

ROS2 is Confusing. Let's Build a Kitchen Instead (Part 1)

A simple, step-by-step guide to understanding the core of ROS2: Nodes, Topics, and Services.

Read it here

The Top 10 ROS2 Commands You Should Know

Your go-to guide for building, testing, and debugging in ROS2.

Read it here

Projects

Prompt Based Auto-Aiming and Tracking System

LLM OpenCV Arduino PyTorch

Built a real-time tracking and aiming system that selects targets with a simple prompt and hits with sub-2° accuracy.

Learn More

Realtime Object removal and seamless video restoration

PyTorch OpenCV VLM CUDA

Created a system that removes unwanted objects from live video and fills in the gaps seamlessly, making the scene look natural.

Learn More

Autonomous Maze Solving Using CV and Inverse Kinematics

MATLAB OpenCV Python URDF

Got a robot to see a maze, solve it, and move through it smoothly with some path planning and smart motion control.

Learn More

Turtlesim "Catch Them All"

ROS2 Python Robotics

Designed a multi-robot ROS2 "catch-em-all" simulation where multiple chaser turtles pursue dynamically spawned, randomly moving targets.

Learn More

Synapse: An AI Accelerated Tutor

LLM Ollama Python Gradio

Developed an AI tutor, that features a multi-agent architecture. It uses a router to delegate tasks, a retriever for information and much more.

Learn More

Indian Sign Language Translator

Computer Vision Tensorflow Python Streamlit

Built a bidirectional Indian Sign Language (ISL) translator, converting signs to text and vice-versa.

Learn More