Akhil Turaga

Working on producing automatic image recognition of neural rosettes from the outputs of agent-based simulations and eventually experimental images

Family Practice Established in 2014 since age 12, Western & Indian styles, 130 students (3 are state champions and national qualifiers), 200 performances (15 national), worked w/ 27 Nonprofit Organizations. Manage Bollywood/Hip-Hop department, marketing, community outreach, public relations, and graphic design

Conducted a literature review of over 30 academic and industry sources to identify and quantify 6 key parameters affecting adhesive performance (e.g., peel force, loop tack load, shear resistance, surface tension). Applied Young-Dupre equations to understand how wetting and adhesion behavior are influenced by processing variables like cooling time, surface energy, and substrate temperature. Led the redesign of procedures and protocols by incorporating mathematical modeling of surface tension and surface energy mismatches, resulting in improved test reliability. Determined key experimental variables such as gun temperature, heat plate cooling time, and application method. Demonstrated that better wetting correlated with higher peel force. Authored RDE-25-0231, a protocol that linked application temperature to peel force outcomes using 64 adhesive film samples under varied thermal conditions. Built and tested samples using enhanced methods such as uniform film pressurization with a metal roller to reduce air bubble artifacts in peel testing. Analyzed failure modes during peel testing and refined the process to improve consistency across adhesives with different thermal activation ranges. Designed a DOE (Design of Experiments) using new materials (Dallas backing film) to evaluate interactions between substrate type, adhesive type, and temperature range. Provided insights into resin procurement by identifying optimal substrates for testing, emphasizing the importance of surface energy differentials and identifying issues such as surfactant clumping that can affect test accuracy.

BioE 80 - Introduction to Bioengineering

BMED 2310 Facilitate hands-on lab sessions, grade reports, and guide students through the iterative design process in a core biomedical engineering course.

Bhasin Lab of Biomedicine Identified over 20 cancer cell types from tested cell clusters. Constructed Deep Belief Network trained on bone marrow gene expression data (70/30). Accuracy score of 87% amongst 10X Genomics datasets ranging around 12,290 cells each. Created Gaussian RBM from scratch, implemented hand-crafted SMOTE function to inflate underrepresented training data.

3D Printing on the Nanoscale Utilized MATLAB to design and implement 25 algorithms for the processing and analysis of nanoscale images, crucial for the development of the low-cost 3D printer project funded by the NSF. Created 20 ground truth datasets using MATLAB, produced 5,000 accurate, high-resolution images for validation of the nanoscale 3D printing process.