Rahima Atiq Mir

Robotics and AI Engineer specializing in Neural Decoding and Computer Vision

Based in: Karachi, SindhCurrently: Robotics Instructor, Roboautomators

A Robotics and AI student with extensive experience in developing neural decoding pipelines, GAN-based image reconstruction, and autonomous robotic systems. She has a proven track record of building complex AI models and instructing the next generation of robotics enthusiasts through hands-on hardware and software integration.

Experience

Robotics Instructor

RoboautomatorsworkSep 2026 – Present

Summary

Designing and executing engaging robotics classes for school-going kids and interested professionals. With the use of arduinos and sensors such as IR, UltraSonic etc, we teach children something new about robotics everyday. Younger kids work with lego kits that include a smart hub and built in sensors. The programming of each robot is done through makerzoid and pictoblox on arduino.

What I did

Teaching children something new about robotics everyday using arduinos and sensors such as IR, UltraSonic etc.
Working with younger kids using lego kits that include a smart hub and built in sensors.
Programming robots through makerzoid and pictoblox on arduino.
Taught robotics as part of school curricula.
Guided students in building robots including dolphin, sweeper bot, electric fans, big hand fans, step man, and crocodile bot.

Results

Taught after-school classes with batches of at least 10 students.

ArduinoIR SensorsUltraSonic SensorsLego KitsMakerzoidPictoblox

Robotics Intern

SystematicHUBinternshipNov 2025 – Dec 2025

Summary

Worked on building a warehouse management system, training the core softwares needed for an obstacle avoidance robot. With an accuracy of 93.21% and the help of PyTorch, OpenCV. Minor work using Jetson Nano, and major work on LLMs, TensorFlow, and various image generators. CV: segmentation, detection, PyTorch, OpenCV, image generators, Jetson Nano

What I did

Trained the core softwares needed for an obstacle avoidance robot.
Performed minor work using Jetson Nano.
Performed major work on LLMs, TensorFlow, and various image generators.
Applied Computer Vision techniques including segmentation and detection.
Created the software to make the warehouse robot workable, including object detection and obstacle avoidance programs.

Results

Achieved an accuracy of 93.21% for the obstacle avoidance robot software.

PyTorchOpenCVJetson NanoLLMsTensorFlowImage GeneratorsSegmentationDetection

AI Intern

Arch TechnologiesinternshipJul 2025 – Aug 2025

Summary

Developed and optimized AI/ML models, contributing to data analysis, training, and cloud deployment. Collaborated with engineers to enhance model performance and documented project outcomes.

What I did

Contributed to data analysis, training, and cloud deployment.
Performed batch size tuning and optimized hardware utilization to speed up execution time per frame for real-time applications.

Results

Achieved model accuracy ranging from 80% to 92%.
Handled datasets of around 1-2GBs.

AIMLCloud DeploymentData Analysis

Skills

Technical

PyTorch

Python

Machine Learning

Computer Vision

Deep Learning

Artificial Neural Networks

Arduino

Natural Language Processing

OpenCV

LLMs

GANs

CNN

Keras

Rhino/Grasshopper

FastAPI

React

TensorFlow

Google Sheets API

Evolutionary Computing

Vite

Jetson Nano

IFTTT

Projects

Final Year Project: NeuroVision

Lead Developeracademic

Developed a CPU-compatible neural decoding pipeline that reconstructs visual stimuli from fMRI-derived brain activity using feature decoding and a pretrained GAN-based image generator. Adapted the research pipeline for Windows environments, integrated it with a FastAPI backend and React-based NeuroVision interface, and implemented an end-to-end workflow for uploading neural data and displaying reconstructed imagery., with an accuracy of 61.12%.

fMRICNNAlexNetCaffeNetGANPyTorchComputer VisionFastAPIReactViteWindows CPU

Mental Health Chatbot

Developerpersonal

Developed a conversational AI chatbot using NLP and LLM-based techniques, building a sentiment analysis model in PyTorch/Keras to generate context-aware responses. With an answer accuracy of 92.1% solving basic mental health problems and providing with temporary assistance.

NLPLLMPyTorchKerasSentiment Analysis

Holographic PC

Designer and Implementeracademic

Designed and implemented a novel agent-based architecture for the desktop assistant, enabling natural language interaction for executing systemlevel tasks. Integrated speech recognition and NLP pipelines to interpret user commands for holographic desktop assistant operations. Integrated speech recognition modules and natural language understanding (NLP) pipelines to interpret user commands for holographic desktop assistant operations.

NLPSpeech RecognitionAgent-based Architecture

Voice to Print System

Engineerpersonal

Engineered a voice-to-3D-model automation workflow, integrating APIs (IFTTT, Google Sheets) with Rhino/Grasshopper to translate spoken instructions into structured geometry updates in real time. Developed a Python-based API integration layer to connect IFTTT and Google Sheets, enabling real-time data synchronization for the voice-to3D-model workflow.

PythonIFTTTGoogle Sheets APIRhinoGrasshopper

Education

Bachelors in Robotics and Artificial Intelligence

SZABIST UniversityRobotics and Artificial IntelligenceSep 2022 – Jun 2026

Thesis. NeuroVision

Coursework

Deep learningArtificial Neural networksMachine learningNatural Language processingComputer visionEvolutionary ComputingResearch Report and Writing

Recognition

Awards

Project "NeuroVision" recognition and presentation

Aga Khan University, CIME SIMPACT2026

Recognition and presentation of the NeuroVision project.

First prize winner at ZAB EFEST

SZABIST University2022

Won first prize for project "Holographic PC".

Social impact & activities

Project "NeuroVision" Presentation

CIME SIMPACT 2026conference talk2026

Presented the NeuroVision project at Aga Khan University.

Hire RahimaReach out about a role, a contract or a conversation.For recruiters

Rahima's twin is AI, it can make mistakes.

More people on Kanvis

Browse more people Create your own Kanvis