Ali Afzoon

•​ Develop an AI agent to help with bug reports. Achieve 70% successful bug root investigation.​ ​• Parse artifacts of the reports, data modeling and database management.​ ​• Design and develop RESTful API to manage the files.​ ​• Design and develop MCP tools to support the AI agent.​ ​• Maintain the codebase, create CI/CD pipelines.​ ​• Complete multiple trainings including Virtuozzo Hybrid Infrastructure training​.

Fine-tuning a bone segmentation model based on the Segment Anything Model using PEFT method. Improved Dice Score and Sensitivity on Vertebrae. Created an efficient data pipeline to speed up the delivery 40x. Tracked and documented the learning process. A collaboration.

Designed, trained and deployed custom Deep networks for time series forecasting on power grid data. Predicted AFRR activation direction with 〜70% accuracy among other tasks. Containerized and deployed models to the server. Created an efficient data pipeline to retrieve data and write to an efficient time-series database in order to serve multiple models in production with different needs. Tracked and compared performance of new models. Student job.