Hi, I am

Dheeraj Kallakuri

About

About Me

Hi, my name is Dheeraj Kallakuri. I completed my Masters in Robotics and Autonomous Systems with specialization in Artificial Intelligence from Arizona State University in May 2024. I am an enthusiastic software developer with over two years of experience at Zeus Learning. At Zeus, I enhanced user engagement by significantly 40% through the deployment of interactive learning modules. At ASU, my passion for robotics is demonstrated by my work at the Battery Electric & Intelligent Vehicle lab. Where I engineered advanced sensing and perception modules for autonomous vehicles. With my proficiency in Python and strong skillset in ROS, Sensor Fusion, and Computer Vision have successfully implemented solutions using TensorFlow and Pytorch. My technical expertise encompasses various programming languages, front-end and back-end development, data engineering, and embedded systems.

From my first year of undergraduate studies, I have been deeply involved in robotics, tackling projects that integrate both hardware and software. I am confident that my contributions often exceed the standard requirements, and I have a strong track record of completing prototypes for various proof-of-concept projects. Throughout my academic career, I approached each project with a product-oriented mindset, handling everything from brainstorming and idea development to marketing. In addition to my coding and algorithmic expertise, I excel in planning, execution, and development. My experience in competitive robotics, coupled with my master's degree in Robotics and AI, has equipped me with advanced skills and knowledge of the latest technologies. I strive to make a meaningful impact by addressing real-world problems.

  • Name: Dheeraj Kallakuri
  • Email: dkallak1@asu.edu; kallakuridheeraj@gmail.com
  • Phone: +1 203-919-1956
  • Profiles: LinkedIn GitHub LeetCode HackerRank Devpost

Download Resume

Education

ASU Logo
2022-2024

Arizona State Univeristy (ASU)

Master of Science (MS) in Robotics and Autonomous Systems (AI)
CGPA: 4.0
MU Logo
2015-2019

University of Mumbai

Bachelor of Engineering (BE) in Computer Engineering
CGPA: 3.57

Experience

Caterpillar Logo
Nov 2024 - Present

System Engineer

Caterpillar, Tucson, Arizona
System Integration and Testing on Autonomous Trucks.

  • Create, validate, and troubleshoot autonomy software systems for autonomous machines, including planning, perception, and positioning technologies.
  • Conduct thorough onboard system validation, execute test plans for autonomous trucks, and analyze test data to improve system performance.
  • Identify and resolve mechanical, electrical, and software issues in autonomous haul trucks, documenting and communicating findings effectively.
  • Prepare detailed test reports with actionable insights and recommendations based on testing outcomes.

    • BELIV Logo
    Feb 2023 - May 2024

    Graduate Student Assistant

    BELIV Lab, ASU, Mesa, Arizona
    Implementation of Computer Vision algorithms in Intelligent Parking System.

  • Deployed the perception module for an Intelligent Parking System, achieved a 95% accuracy rate in detecting empty parking spots using a custom YOLOv5 model and ensuring seamless system integration of hardware and software components.
  • Engineered advanced sensing and perception modules of 2022 Mustang Mach-E autonomous car, enhancing sensor fusion capabilities by 30%.
  • Developed an effective test harness and debugging methodology of sensors and software modules for the ROSMASTER X3 bot, reducing troubleshooting time by 40%.
  • Researched and integrated a virtual setup in CARLA and Autoware for testing ADAS vehicles, improving testing efficiency by 25%.

    • Zeus Learning Logo
    Jul 2019 - May 2022

    Software Engineer

    Zeus Learning, Mumbai, India
    Web developement on e-learning websites.

  • Programmed and deployed interactive end-to-end learning modules and enhanced user engagement by 40% using Dotnet and Angular.
  • Designed and implemented frontend components for data visualization and report generation using the latest frameworks and improving website efficiency by 35% using React.
  • Created and optimized algorithms for mathematical questions, increasing accuracy and reliability by 80% using Typescript.
  • Trained and mentored 7 junior developers, boosting team performance by 45% and ensuring project progress through regular client feedback sessions, leading to a 30% increase in client satisfaction.

    • Central Railways Logo
    Sept 2020 - Feb 2021

    Freelance Project Coordinator

    Central Railways, Mumbai, India
    Mentoring safety and maintenance automation projects.

  • Prototyped 5 railway safety and maintenance automation projects using computer vision, machine learning, and IoT, increasing automation by 60%
  • Tested and deployed projects on embedded devices like Raspberry Pi and Arduino, improving operational efficiency from 30% to 65%
  • Interacted with railway officials to demonstrate project functionality, successfully convincing them of an 80% reduction in human resource requirements and time for maintenance and routine tasks.

    • Hackathons

    sunhacks Logo
    Sept 2024

    Sunhacks 2024

     Sunhacks Club, ASU, AZ
    Sparky - The Hardware Guru

  • Designed a fully functional chatbot with voice control, enabling natural interactions and answering questions about hardware and electronics.
  • Integrated the Gemini API to provide accurate and context-aware responses to technical inquiries.
  • Enhanced the chatbot with a detection feature to identify and address questions about unknown tools or electronics.
  • Incorporated a custom vision model into the chatbot to support advanced capabilities, including object or tool recognition.
  • Successfully executed hardware and software integration on a Raspberry Pi platform, showcasing end-to-end functionality in a compact system.

    • ohack Logo
    Oct 2024

    O-Hack 2024

     Opportunity Hack, ASU, AZ
    Alpaca3 - Custom GPT Solution for Liminal Works

  • Built a custom chatbot using RAG (Retrieval-Augmented Generation) architecture with a Flask backend and a Streamlit frontend.
  • Created a MongoDB database to store user queries and responses, with a system to update the knowledge base periodically.
  • Designed a content generator to produce PDFs with Q&A responses, simplifying actionable suggestions for Liminal Works volunteers.
  • Integrated OpenAI and Gemini APIs for generating accurate query responses.
  • Presented progress and discussed scalability and future enhancements in a productive debrief meeting with Diego González-Medina from Liminal Works.

    • soda Logo
    Oct 2024

    Hack SoDA 2024

     SoDA Club, ASU, AZ
    GenderBiasGPT for Regional Languages 🥇

  • Developed a robust backend using Flask to support efficient system functionality.
  • Created an interactive and user-friendly UI with Streamlit, ensuring a seamless user experience.
  • Integrated backend and frontend components to build a cohesive and well-functioning system.
  • Conducted research and testing on embeddings and tokenization for Hindi text to enable effective Hindi language processing.
  • Contributed to building an accessible tool for detecting gender bias in regional languages, with a focus on Hindi.

    • Skills

    Programming Languages
    PythonC/C++
    Robotics
    IoTEmbedded C AtmelAVRArduino Raspberry PiROS2 MatlabRviz OpenCVNvidia Boards TfLiteTinyML MechatronicsSimulinkSensors Fusion
    Data Engineering
    TensorflowPytorch NumpyPandas Exploration Data AnalysisML Foundations MathplotlibData Visualization Statistical AnalysisPredictive ModelingData Mining and Analysis
    Generative AI
    ChatGPTLlama Quantised Llama(O-llama) Gemini Bidirectional Encoder Representations from Transformers(BERT) Generative Adversarial Networks(GAN)Transformer-based models Retrieval-Augmented Generation(RAG)Hugging Face Kaggle datasetsPretrained models Large Language Models(LLMs)LLMs API Google Cloud AIVector Embedding Vector DatabaseVertex Custom AI Models/AgentsFine tuningPipelines Prompt EngineeringSummarization
    Web Stack
    React.JSJavaScript TypeORMGraphQL TypeScriptHTML CSSJSON C#DotnetAngular FlaskStreamlitWeb API Web Hosting
    Tools, Databases, and OS
    DockerMariaDB MongoDBPostgreSQL GitSQL NoSQLWindows LinuxmacOS JiraTrello MiroGPU CUDATensorRT Google ColabJupyter Notebook vscodeAnaconda
    Concepts
    Artificial IntelligenceMachine Learning Reinforcement LearningImage Processing Transfer LearningData Strucure and Algorithms Object DetectionComputer Vision Perception in RoboticsSystem Engineering Product DevelopmentPrototyping Embedded Machine LearningNeural Networks Deep LearningHuman-Robot Collaboration Markov Decision ProcessLarge Language Models Generative AI

    Research Papers

    springer logo
    July 2023

    M-Lens (IoT-based deep learning device)

    Published in Internet of Things (IoT): Key Digital Trends Shaping the Future in Springer.

    The quality of instruments used in the manufacturing industry is crucial. Ensuing fatigue and inattentiveness limit the reliability of manual inspection. M-Lens is a hand-held device that automatically classifies defective parts even if the defects are novel and previously uncategorized. This allows inspection workers to quickly identify not only previously known defects with great accuracy but also novel defects by adding new/custom defects to the cloud-based defect repository and retraining the deep learning model on the server. The concept of transfer learning is used to enable a shorter training time. The device is made using off-the-shelf Raspberry Pi and camera extensions. It is a beneficial solution where production ecosystems are constantly changing, and time is invaluable to the business. This paper discusses the major features of implementation, working, and analytical results of the final device.

    Link to paper

    ieee logo
    March 2022

    Real-time Alert System for Auxiliary Transformer Failures

    Published paper in IEEE Xplore.

    Auxiliary transformers are used extensively in railway systems. Indian Railways use them primarily to convert the 25-kilo volt AC supply into 230-240-volt AC supply as required for various applications. Due to their varied use, many of these auxiliary transformers are often positioned in remote locations. In case of failure of AT supply, no proper feedback system exists as of now. Due to this the lead time on failure attention increases rapidly depending on the location of the transformer. Sometimes this may result in signal failure leading to an increase in train traffic. To cut down on human effort and cost involved, as well as provide a system that continuously monitors these auxiliary transformers, an automated system based on GSM technology is suggested. This system periodically provides updates and generates immediate alerts in case of the occurrence of the failure of an auxiliary transformer, hence bringing a drastic reduction on lead time for failure attention where a failure escalates into costly service losses.

    Link to paper

    research-gate logo
    June 2021

    Implementation of Implantation-Stagger Measuring Unit

    Published in the International Journal of Engineering and Management Research (IJEMR), Vol-11 issue-3.

    The electrical traction system of railways is a combination of physical upright structures and OCL(Overhead Contact Lines). The horizontal distance from the center of the track to the OHE mast is called implantation, the horizontal displacement of overhead contact wire concerning the center of the railway track called stagger, and the perpendicular height of overhead contact wire from the ground is periodically checked by a lineman to ensure a safe distance from the railway track. In this paper, we have put forth the idea of building a distance-measuring device to measure the implantation and stagger without touching the objects using Open CV on Raspberry Pi with a camera module which will be placed a the center of the track. The system will have two features. To measure the distance of the nearest poles, the camera has to be placed facing the mast perpendicular to a circle of diameter appropriate which is placed on the pole for measurement purposes. To measure the stagger, the camera has to be placed facing the overhead wire from the center of the track.

    Link to paper

    Projects

    Projects

    Intelligent Parking System

    Dynamic Navigation to the nearest parking spot from a current location.

    Python, Computer Vision, Jetson Orin, MongoDB

    NHTSA Image

    NHTSA Collision Dataset Analysis

    Study and analyze the ADS and ADAS Level 2 collision and summarize the trends.

    Python, Data Visualization, Predictive Modeling

    Posture Correction Chair

    Chair providing real-time feedback for users to maintain good posture.

    Python, C++, Arduino, Reinforcement Learning

    Hand Gesture Recognition System

    One Stop Shop: Hand gesture recognition apps.

    Python, Mediapipe, Raspberry Pi, Machine Learning

    High-Accuracy Keyword Spotting on Edge

    Real-time audio analysis embedded system to detect specific absolutist keywords.

    C++, Arduino, Tensorflow Lite

    M-Lens Deep Learning device

    Handheld defect detection device for Airplane Industry.

    Python, Computer Vision, Transfer Learning, Raspberry Pi, AWS

    Mini Autonomous Car

    Autonomously navigate by following a line and responding to specific traffic signs.

    ROS2, Jetson, Open CV, Classification

    PiBot: Rescue Surveillance Robot

    A server controlled Search and Rescue (SAR) operations robot.

    Python, Sensor Fusion, Raspberry Pi, Web App

    Autonomous Obstacle Detector Bot

    An autonomous robot capable of detecting and avoiding obstacles.

    C++, Arduino, Sensor Fusion

    SONAR Logger

    The system logs all the obstacles in the 180-degree range.

    Matlab, Simulink, Arduino, Data Processing

    Sparky: Hardware Guru

    An interactive assistant that educates users about various tools & components.

    MultiModal LLM, Raspberry Pi, Speech Recognition, Yolov8

    Alpaca 3: CustomGPT

    An AI assistant for Liminal Works Organization to understand concepts of Theories of Change (ToC).

    LLM API, Chatbot,PDF Generation, MongoDB, RAG, Flask, Streamlit

    Regional GenderBiasGPT

    Developed a tool to detect gender bias in Vietnamese and Hindi text inputs.

    Transformers, Cosine Similarity, Tokenization & Embedding, Streamlit

    Image Caption

    Colorizer

    Developed a web app using Colorful Image Colorization technique to convert B/W to Color images.

    Python, CNN, Unsplash API, Streamlit

    QR Code App

    Name-o-Color

    This app generates a unique color based on a user's name.

    Python, Streamlit

    youtuber analysis

    YouTuber Analytics

    Tool that provides detailed insights of YouTubers using Quantised Llama and Youtube API.

    Gen AI, Quantization, Llama, Hugging face, Langchain

    Imdb Review Classifier

    Imdb Review Classifier

    Fine-tuned BERT(NLP model) for sentiment analysis to classify movie reviews as bad or good.

    Gen AI, Open AI, Chatbot, GUI, LLM, Python

    Pdf Chatbot

    Pdf Chatbot

    Developed a Q/A chatbot using RAG architecture to answer questions from analyzed PDF content.

    GenAI, RAG, Chatbot, LLM, Langchain, Python

    Image Caption

    Image Summarize

    Built an Image Caption model that summarizes the given input image.

    ResNet-50, LSTM, GPT Encoder, Kaggle, Python

    QR Code App

    All-in-One Filters

    Single destination for all photo filters, including iPhone, Instagram, and basic options.

    OpenCV, Image processing, Image Filters, GUI

    transformer architecture

    Scratch GPT

    Developed a decoder-only transformer LLM from scratch by following Andrej Karpathy's tutorial.

    Gen AI, LLM, Transformers, GPT-2 Architecture, Python

    chatbot

    MiniChatBot GPT-3.5

    Real-time chatbot, allowing users to interact with GPT-3.5 Turbo.

    Gen AI, Open AI, Chatbot, GUI, LLM, Python

    Joke App Logo

    Joke App

    Built a joke app that displays and reads jokes and GIFs.

    API, TTS, Threading, GUI, Python

    Images to pdf

    Images to pdf Conversion

    Created a python app that combines keyword-specific images to PDF document.

    API, Document Conversion, Threading, GUI, Python

    QR Code App

    QR Code Generator

    Designed a QR code generator app that creates customizable QR codes for validated URLs.

    Validation, Python, QR Code, Custom, GUI

    embedded ML

    Embedded Machine Learning Projects

    Implemented 6 mini projects during course work.

    C++, Arduino, Neural Networks, Sensor Fusion

    Railways

    Central Railway Projects

    Deployed 3 POC projects for Safety and Maintenance Department.

    C++, Arduino, Image Processing, Sensor Fusion

    robotic arm

    Robotic Arm Projects

    Executed 3 specific tasks using UR robots and Transporter Networks.

    UR programming, Python, Simulation, CLIPort

    makeblock

    Makeblock Projects

    Launched 2 innovative projects using educational robots.

    C++, Arduino, Sensor Fusion, Communication

    certificate

    Gen AI Certification Projects

    Developed 2 projects during Nvidia, and Google certification courses.

    Python, Neural Networks, Gen AI, Jupyter Notebook

    Certifications

    Certifications

    google certification
    universal robots certification
    matlab certification
    microsoft certification
    nvidia certification
    aws certification
    google certification
    Qualcomm courses
    microchip courses
    coursera courses
    Scrimba courses

    Awards

    Awards

    award 1
    award 2
    award 3
    award 4
    award 5
    award 6
    award 7
    award 8
    award 9
    award 10
    award 11

    Contact

    Contact Me

    Location

    Glendale, AZ 85303