Parnian Ghapandarkashani Email & Phone Number

United States

Los Angeles, California, United States

• Selected from over 3,000 applicants for the prestigious Break Through Tech AI program at UCLA.Completed advanced Machine Learning coursework designed by Cornell and UCLA faculty, with a focus on foundational and applied AI concepts.• Collaborated on a Vision Language Model (VLM) project in partnership with Qualcomm, fine-tuning Microsoft’s Florence-2 model for weapon detection using the LoRA (Low-Rank Adaptation) technique. Leveraged a custom dataset from Roboflow to enhance public surveillance and security applications, achieving a Mean Average Precision (mAP) score of 0.63.

Los Angeles, California, United States

• Collaborating with UIUC and USC on a joint project called Universum, aimed at advancing machine learning for human perception and neural interfaces.• Developing a comprehensive machine learning model to understand and predict human perception using EEG data from commercially available headsets.• Designing a hierarchical framework for each aspect of perception, employing curriculum learning to train models on tasks of varying complexity.• Generating perceptual embeddings to enhance the performance of generative AI models and enable applications such as conditioning diffusion models, prosthetics, and virtual reality.• Contributing to model development at UCLA, with a focus on building advanced machine learning architectures rather than data collection.

Los Angeles, California, United States

As a mentor for the 'Introduction to Electrical Engineering' course at UCLA, I provide both practical lab support and theoretical instruction. This includes guiding students through the use of lab tools such as oscilloscope, breadboarding, debugging circuits, and writing code for microcontrollers, as well as explaining fundamental concepts like circuit analysis, sensor fusion, PID control principles, etc. My role is to help students develop both hands-on skills and a solid understanding of electrical engineering concepts and lab skills.

Los Angeles

• Conducting research on Generative AI and diffusion models for virtual histology, with a focus on real-time applications during surgery.• Leading the data preparation phase, including registering H&E image modality (RGB) to FLIM (Fluorescence Lifetime Imaging Microscopy) data using the WSIReg Python library and applying non-uniform fast Fourier transform (NUFFT) techniques for FLIM data interpolation prior to model training, ensuring high quality ground truth data.

Los Angeles, California, United States

We've developed an innovative, user-friendly GUI platform to streamline the control of "ferrobots" utilizing microfluidic droplet manipulation technology. This technology automates the process of COVID-19 tests, where ferrobots (small magnetic robots) carry droplets over a PCB navigation floor. By replacing manual, spreadsheet-based path input with an intuitive user interface, we are hiding the low-level details of microcontrollers with a user-friendly GUI where users can draw path for ferrobots to move on the PCB.

Bachelor Of Science - Bs, Computer Engineering, 3.93

Activities and Societies: Machine Learning Researcher, CruX neurotech organization @ UCLA Computer Vision Researcher, Ozcan Research Group.

High School Diploma, Mathematics & Physics (Theoretical Branch), 19.97/20