Preethi Amasa

Software Engineer - AI at The Citizen Project (2026-07 – Present)

Ship features on a production Next.js/AWS Amplify application within an Agile/Scrum team, coordinating via Slack and Confluence.
Lead the rollout of AI coding agents into CI/CD and QA workflows to streamline the engineering process.
Manually test and document each feature and ticket fix prior to release, and support SRE work by triaging production issues and auditing environment configs.

AI Research Engineer, Vision-Language Models at South Dakota State University (2025-08 – 2026-06)

Designed RegiGrow, a continual learning framework combining Mixture-of-Experts LoRA adapters with an automated, entropy-based growth mechanism that lets the model expand its own capacity as needed rather than on a fixed schedule.
Built the training infrastructure supporting this research, including automated pipelines that processed 45,000+ video frames across HPC GPU clusters to fine-tune a frozen ImageBind embedding encoder.
Evaluated the framework on the ShanghaiTech, UCF-Crime, and XD-Violence video anomaly and violence-detection benchmarks to validate performance against established baselines.
Applied Claude Code and Cursor for pipeline development, prompt iteration, and test generation, reducing development time on training scripts.

Software Engineer, Backend and Full Stack at South Dakota State University IT Division (2025-01 – 2025-08)

Led end-to-end development of 15+ production REST API features and web/console applications, from requirements and design through deployment, using C#, ASP.NET, Blazor, and SQL Server, supporting 1,000+ active users.
Built dynamic data collection and reporting workflows via the Formsite and SharePoint APIs, and automated backend pipelines with iPaaS for scheduled and event-driven processes.
Optimized SQL queries, joins, and CTEs to cut execution time by 40%, and introduced code-review validation improvements that reduced production errors by 30%.
Manually tested and documented each feature and ticket resolution, producing technical documentation for codebases, APIs, and workflows that reduced onboarding time, while providing ongoing troubleshooting support for deployed systems.

Data Analyst at Prowork (2023-11 – 2024-04)

Built automated SQL and Python pipelines for reporting processes previously performed manually, reducing manual effort by 40%.
Performed feature engineering and statistical analysis using Pandas and NumPy across 20+ stakeholder reporting requests.

About