Profile Photo

Harshaa Hiremagalur Narayan

Software Developer

linkedin

Contact

Address
San Jose, California

Education

2023
Master's in Computer Science
Rensselaer Polytechnic Institute
2021
Bachelor's of Engineering
MS Ramaiah Institute of Technology

Expertise

Python
HTML, CSS, JavaScript
Git, Github Actions
AWS
Windows App Development
SQL, PostgreSQL

Languages

English

Kannada

Hindi

Tamil

This page has been viewed ___ times

Experience

Jan 2024-Present

Velmeni Inc. | Remote

Software Engineer

1. Spearheaded the integration of PACS servers with the company's AI driven report generation module using core Python, delivering the company’s first live project.
2. Secured $80,000 in recurring annual revenue from a single client by deploying the integrated PACS solution.
3. Engineered a high-performance syncing service for dental imaging software using Python, Django, Celery, AWS (S3, Lambda, SQS), and PowerShell, streamlining data workflows.
4. Achieved $100,000 in cost savings by developing the syncing service in-house, replacing third-party tools and enhancing operational control.

May 2022 - Aug 2022

IBM | Almaden,San Jose

Graduate Extern (Internship)

1. Explored Federated Unlearning with IBM.
2. Used PyTorch and Matplotlib for model customization and visualization.
3. Implemented Grid Search for hyperparameter tuning.
4. Built a Python-based evaluation method for Federated Unlearning.

projects

Dec 2024-Present

Cloud based collaborative document editor

1. Development of a collaborative document editing platform using vanilla JavaScript for the frontend and real-time interactions.
2. Integration of Quill as the rich text editor and ShareDB for real-time synchronization and conflict resolution.
3. Backend built using AWS Lambda, API gateway, AWS cognito with DynamoDB for efficient and scalable document storage.
4. Implementation of simultaneous multi-user editing.
5. Utilization of WebSocket for real-time collaboration, providing a seamless user experience.

Mar 2023-Apr 2023

Implementing k-medoids clustering in parallel

1. Wrote parallel implementations of k-medoid clustering components using CUDA and MPI.
2. Wrote a parallel implementation of silhouette coefficient using MPI.
3. Used rice gene expression dataset to evaluate the implementation. Achieved highest silhouette score of 0.65.
4. Achieved a significant speedup of 7x for clustering and 20x for evaluation compared to serial implementation.