👋 Hello, I'm

Pavan H Bhakta

Exploring the intersection of Algorithms and Amazon's Ecosystem

An in-depth analysis of how data structures and algorithms power Amazon's global services, from e-commerce recommendations to cloud computing infrastructure.

Explore Case Studies Get in Touch
Scroll to explore
Algorithms Data Structures AWS E-commerce Machine Learning

About Me

Pavan H Bhakta

Aspiring Software Engineer

Portfolio Details

Course Algorithmic Problem Solving
SRN 01FE22BCS175
Instructor Prakash Hegade
University KLE Technological University
Domain Amazon

My Journey

I am a passionate technologist with a deep interest in solving complex problems through elegant algorithms and efficient data structures. My academic journey at KLE Technological University has equipped me with strong fundamentals in computer science, while my personal projects have allowed me to apply these concepts to real-world challenges.

I'm particularly fascinated by how companies like Amazon leverage advanced algorithms to create seamless experiences at massive scale. This portfolio explores the intersection of theoretical computer science and practical business applications, demonstrating how algorithmic thinking drives innovation in modern tech ecosystems.

Technical Expertise

Programming Languages

C++ Python JavaScript Java

Web Technologies

React.js Node.js HTML/CSS REST APIs Express.js MongoDB

Education

Bachelor of Engineering in Computer Science

KLE Technological University

2022 - Present

Pre University Education(Class 11 & 12)

Saraswati PU College

2020 - 2022

Schooling Class(1-10)

Marthoma Central School Honavar

2010-2020

“I knew that if I failed I wouldn't regret that, but I knew the one thing I might regret is not trying.”

Steve Jobs — Jeff Bezos

Featured Projects

Applying programming concepts to solve complex challenges across various technical domains

Computer Vision

Traffic Sign Detection Model

Developed an advanced GUI-based traffic sign detection system leveraging the YOLOv11 architecture for real-time object detection and classification. The system integrates Python's Tkinter library to provide an intuitive user interface, enabling users to stream live video for traffic sign recognition. Additionally, Python's text-to-speech (TTS) library was utilized to announce detected traffic signs audibly, enhancing accessibility and usability. ensuring high accuracy and reliability in real-world scenarios.

PyTorch Python-Tkinter YOLO CNN OpenCV Transfer Learning
Achieved 94% accuracy on the Indian Traffic Sign Recognition Benchmark
Designed a real-time voice feedback system using Python’s text-to-speech (TTS) libraries for auditory alerts
Developed a Python-based Tkinter GUI for real-time video feed processing and dynamic traffic sign visualization
View Project
Full Stack Web Development

QuickEats – Food Delivery Platform

Designed and developed a robust food delivery platform using the MERN stack, enabling users to browse menus, place orders, and track deliveries in real-time with seamless user experience. The application incorporates role-based access control, admins can manage restaurant listings and orders, while customers can view their order history and receive personalized recommendations based on their preferences.

MongoDB Express React Node.js
Implemented secure user authentication for both customers and admins using JWT-based authentication
Built a separate Admin Panel for managing menu items, adding or removing items, and updating order statuses.
Integrated Cloudinary for efficient image storage and retrieval, enabling dynamic menu updates with images.
View Project
Computer Architecture

3-Address Indirect Mode Processor

Designed and simulated a complete 16-bit processor architecture using Logisim that implements a three-address indirect addressing mode, enabling more efficient memory access patterns. The design features a 5-stage pipeline with hazard detection, forwarding units, and branch prediction to minimize stalls and maximize throughput in instruction execution.

Logisim Digital Logic Pipeline Architecture Assembly
Implemented a full RISC instruction set with 24 operations
Designed a 5-stage pipeline with hazard detection and forwarding units
View Project
NVIDIA Project Intern

Virtual Hardware Simulation & Code Analysis

Developed a bytecode analyzer for Hardware Sequence Language (HSL) to inspect I2C communication and GPIO operations. Developed the entire tool in C++ with efficient memory handling and performance-focused design

C++ Data Structures Algorithms
Created a simulation module that takes in HSL bytecode, allocates memory, and acts as a virtual device, providing output like an HSL simulator
Designed a memory operations table for streamlined memory lookup
Designed an optimization module to merge consecutive I2C writes and overlap programming delays, improving execution efficiency achieving 30-45% performance improvements
View Project
NVIDIA Project Intern

LLM Paired Comparison Experiment

Using an LLM to compare image sharpness through paired comparison using prompt engineering approach that enables an LLM to actively participate in a paired comparison experiment, where it evaluates and compares the quality of image pair

Python Hugging Face Gemini-api
Developed a Preference Matrix & JND Results – A structured matrix ranking image sharpness and identifying Just Noticeable Difference (JND) levels.
Designed a prompt engineering methodology enabling an LLM to actively engage in paired comparison experiments, evaluating and ranking image quality with precision and consistency.
In Progress
View Research

For More Projects and Contributions check out my Github :)

Github