Harsh Thakkar

Software Engineer
Bhavnagar, IN.

About

Passionate software engineer with strong foundations in algorithms, data structures, and object-oriented design. Proficient in building scalable distributed systems and backend services using C, C++, Java, and Python. Combines expertise in cloud technologies and DevOps, leveraging AWS, Kubernetes, and Terraform with a proven track record in developing fault-tolerant solutions that deliver measurable performance gains. Eager to drive innovation and solve complex problems at scale, while optimizing deployments and enhancing user experiences.

Education

Dhirubhai Ambani University (DAU)

Masters of Technology

ICT (Software Systems)

Grade: 9.00

Government Engineering College

Bachelor of Technology

Information Technology

Grade: 8.77

Sir Bhavsinhji Polytechnic Institute

Diploma

Information Technology

Grade: 9.84

Skills

Programming Languages

C, C++, Java, Python.

Core Computer Science

Object-Oriented Design, Data Structures & Algorithms, Complexity Analysis, Problem Solving.

Distributed Systems & Networking

Unix/Linux, TCP/IP, Socket Programming, RESTful APIs, Microservices Architecture.

Cloud and DevOps

AWS, Docker, Kubernetes, Terraform, Ansible.

CICD and Automation

Jenkins, GitHub Actions, Maven, SonarCube, Trivy.

Monitoring and Logging

Prometheus, Grafana.

Scripting and OS

Bash Scripting.

Version Control

Git, GitHub.

Databases

DynamoDB, PostgreSQL, MongoDB.

Machine Learning & AI

LLM integration and development of intelligent applications.

Interests

Coding
Building new products
Gaming
Movies

Projects

Distributed Load Balancer

Summary

Engineered a distributed load balancer using a Flask-based microservice architecture integrated with Redis for centralized logging and health monitoring, ensuring high availability and fault tolerance. Designed a real-time dashboard to monitor system logs and control service operations, demonstrating capacity to manage and visualize system performance under load. Implemented a round-robin scheduling mechanism for diverse computational tasks (e.g., prime checking, Fibonacci calculation) across multiple server controllers, showcasing expertise in scalable task distribution and system resilience. Integrated automated health checks and performance monitoring routines that optimize system uptime and streamline dynamic resource management.

Personalized Roadmap Generator

Summary

Developed an interactive Python dashboard that leverages user self-assessment data to generate personalized learning roadmaps, aligning with end-to-end solution design from problem definition to deployment. Integrated the Google Gemini (Gemini 1.5) model for dynamic subtopic generation, which provided intelligent recommendations to enhance user learning experiences at scale. Utilized Neo4j graph database technology for efficient storage and rapid retrieval of roadmap data, ensuring system scalability and maintainability under large-scale usage. Containerized the application with Docker to promote consistent deployment across environments and foster agile, continuous delivery practices.

Chat Application Deployment using Minikube

Summary

Developed a real-time chat application using React and Node.js with a MongoDB backend to support 100+ concurrent users, emphasizing robust system design and user scalability. Containerized the solution using Docker and orchestrated deployment on a Kubernetes (Minikube) cluster, implementing pod auto-scaling and persistent storage strategies to handle variable traffic efficiently. Enforced industry best practices in cloud-native development by automating CI/CD pipelines, significantly reducing deployment errors while accelerating update cycles in an agile environment. Demonstrated hands-on expertise in building resilient systems by monitoring and rapidly mitigating 3x traffic spike scenarios through proactive scaling and configuration management.

AlgoLens - Interactive Algorithm Learning Platform

Summary

Designed and built a web-based platform for algorithm visualization using modern web technologies, facilitating understanding of complex algorithms through interactive, real-time execution displays. Developed a dynamic visualization engine capable of running 10+ algorithms, which improved comprehension and debugging efficiency by streamlining manual tracing processes. Implemented custom input handling and step-by-step walkthroughs to simulate algorithm execution, emphasizing your strengths in both problem-solving and user-centric design. Increased engagement among 75+ users by creating an intuitive learning environment that reduces cognitive load and accelerates the learning curve, reflecting innovative application of distributed systems and real-time data management.

LSM Tree Compaction Optimization

Summary

Pioneering research to optimize Log-Structured Merge (LSM) tree compaction, achieving a 20-30% improvement in key-value storage performance at scale. Designing and implementing hybrid leveling and tiering strategies that significantly reduce write (by 30%), read (by 25%), and space amplification (by 40%), ensuring efficient resource utilization in large distributed systems. Conducting in-depth analysis of compaction triggers, data layout, and data movement to minimize query latency and enhance overall system throughput in distributed databases. Focusing on storage efficiency improvements and minimizing write stalls, thereby enhancing the scalability and resilience of databases managing petabyte-scale workloads. Leveraging advanced benchmarking tools to validate optimization techniques, achieving up to a 50% reduction in compaction overhead and demonstrating a strong aptitude for performance tuning and system design.