Venkatakrishnan Logganesh

Dynamic and innovative professional with a proven track record at Stellantis, specializing in deep learning solutions. Successfully automated a CNN-based model to classify automobile images, resulting in a 30% increase in efficiency. Proficient in Python and TensorFlow, with strong analytical abilities that enhance data processing and security compliance across various platforms. Career goals include leveraging expertise to drive further advancements in artificial intelligence applications.

ACCENTURE UNIVERSITY INNOVATION CHALLENGE - CYBERSECURITY(Jan 25) Developed QRguard, an API for QR code analysis with scam detection features, and proposed integration strategies with Scam Shield and third-party platforms. 

Capstone project(Sep 24 - present) - With Boeing Design a comprehensive Battery Management System for airlines operation control centers in Southeast Asia. DETR(Detection Transformer) Table Structure Recognition - Developed a table-structure recognition model for the pubtabnet dataset consisting of around 500000 images. Sentimental Analysis (NLP) Project Imported the BERT backbone, redesigned it with attention masking mechanisms in torch, torchtext and trained it on the IMDB dataset containing 50000 movie reviews for identification and detection of positive and negative sentiments. Medical Image Recognition Binary Classifier (Melanoma Classifier) 01/01/24 to 04/01/24 -  Employed a pre-trained Densenet-201/MobileNetV2 model on PyTorch for melanoma detection using the HAM10000 dataset. Medical Image Recognition Binary Classifier (Skin Cancer) - Utilized a pre-trained triplet neural network from PyTorch to predict skin disease presence using the ISIC2018 dataset. Machine Learning model for sentimental analysis - Built sentiment analysis systems for Spanish and Russian using second order Hidden Markov Models (HMM) in Python. Machine Learning model for food security index -  Developed a linear model to analyze factors influencing the food security index.

Nationality: Singaporean