PerceptiLabs
Machine Learning Intern | Sept 2020 - Nov 2020 | Redwood City, CA
I developed 3 Deep Learning models on the company’s Modeling tool: NVIDIA’s End-to-End self-driving car, American Sign Language Detection, and Machine Fault Detection. Achieved losses ranging from a high of 6% to a low of 0.28%. Actively participated with project supervisor/CTO to add 2 important features in the tool: Principal Component Analysis, and Preprocessing pipeline that would enable users to manage their data in the tool itself rather than using an external application. Cross-collaborated and managed projects with the marketing team to publish all 3 models on the company’s platform: GitHub/YouTube.
SimTech
Data Analyst Intern | Jan 2020 - April 2020 | OneNorth, Singapore
I developed in Python, a Tool Condition Monitoring System, using Keras to predict if a machine (Conveyer Belt) is in good or bad shape post an experiment. Obtained accuracies ranging from 61.5% to 99.8% (Training and Test). Conducted comprehensive analysis of the dataset collected from the Tool lab, with data preprocessing and built-in accuracy checks to catch any irregularities in the vibration of the machine. (Dataset had 9800 sample points with 58 features). Automated program in python. Manual effort reduction from 7 scripts to 1 program.
