Hi, I'm Prakruthi Koteshwar
I’m a software developer with 3+ years of experience building web applications and backend systems. My work has involved full-stack development, cloud platforms like AWS, and integrating AI-based tools into applications. I enjoy building practical solutions, improving system performance, and working on problems that require both engineering and creative thinking.
View My WorkProjects
AI & Machine Learning
Traceability Engine: AI-Powered VCS Automation
CLI tool that prevents documentation drift by synchronizing source code and method documentation using AST parsing and Llama-3.3-70B. Processes 100+ functions in parallel using concurrent execution, reducing manual documentation effort by ~70%.
MNIST Digit Classifier — CNN vs MLP
Built a deep learning system to recognize handwritten digits (0 - 9) from uploaded images. Implemented and compared MLP and CNN architectures, demonstrating CNN’s superior performance for image data. The system includes image preprocessing with OpenCV and a FastAPI-based REST API for real-time predictions.
FinTune-Sentinel: LoRA-Enhanced Auditor
A specialized financial LLM fine-tuned using LoRA (Low-Rank Adaptation) to map 10-K risk factors with high precision. By freezing the base model and training rank-decomposition matrices, I achieved state-of-the-art FinQA accuracy while reducing trainable parameters by over 90%.
Agentic Corrective RAG (CRAG)
In Progress
SupplyChain-Sense: Multi-Tier Risk Mapping
In Progress
Full Stack Development
WeCureIT
Designed and developed a web-based system for a multi-location medical clinic that enables patients to book appointments with doctors based on specialty and facility availability. The platform supports dynamic doctor scheduling, facility capacity management, secure patient/doctor accounts with 2FA, and an admin dashboard for managing clinics, doctors, and website content.
High-Performance Distributed Job Scheduler (Go + gRPC)
Built a distributed task scheduler in Go with a gRPC-based communication layer between a centralized Master node and multiple Worker nodes. Implemented a priority-aware scheduling engine using a bin-packing strategy to efficiently allocate resources across the cluster and maximize utilization. Added constraint-based scheduling (Node Affinity) to intelligently match tasks with appropriate hardware profiles. The system is designed for scalability and resilience, with ongoing enhancements including fault recovery, resource quotas, and real-time cluster telemetry.
Blog Platform
Developed full-stack blogging platform with markdown editor, user authentication, and SEO optimization. Implements lazy loading for performance.
Logic & Programming
Tower of Hanoi Solver
Dijkstra's Shortest Path
Created interactive simulation of Dijkstra's algorithm for finding shortest paths in weighted graphs. Includes step-by-step visualization and performance metrics.
Password Organizer using Tree
Passwords are stored and organized using a tree structure that enables efficient searching, insertion, and categorization of credentials by platform or service.
Experience
Software Engineer Consultant
- Evaluated and implemented cloud-based High Performance Computing (HPC) infrastructure on AWS, focusing on scalable and cost-efficient workload execution.
- Configured Linux-based environments, shared storage systems, and the Slurm job scheduler for distributed HPC workloads.
- Provisioned AWS Spot Instances with autoscaling and monitoring using CloudWatch and AWS Budgets to optimize compute costs and resource utilization.
Graduate Research Assistant - George Washington University
- Developed full-stack research tools using Node.js and React (Remix) to support web security and privacy data analysis.
- Integrated Google Gemini AI APIs to automate data processing, content analysis, and intelligent query workflows.
- Implemented Elasticsearch indexing and search pipelines to enable fast retrieval and analysis of large research datasets.
- Built web scraping pipelines to collect and structure web data for security and privacy research experiments.
Software Intern
- Developed UiPath-based RPA workflows to automate data processing tasks in finance applications, reducing manual effort and improving operational efficiency.
Software Developement Engineer in Test - II
- Developed GenAI-powered testing tools with Node.js backend services and modern frontend interfaces to automate test case generation and workflow analysis.
- Integrated LLM-based capabilities into enterprise testing frameworks to support intelligent test generation and validation across applications.
- Built and enhanced automation utilities for Java-based applications, enabling scalable testing and AI-assisted debugging workflows.
- Designed full-stack solutions combining frontend interfaces, Node.js services, and AI models to improve testing efficiency and developer productivity.
Resume
Education
Master of Science in Computer Science
George Washington University
Graduation: 2026
GPA: 3.8/4.0
Skills
Languages:
Tools & Frameworks:
Certifications
Google Cloud Professional
Issued: March 2024
AWS Cloud Foundation
Issued: Aug 2024