MANEESH SINGH

At OutperformAI, I engineered a cutting-edge Stable Diffusion model, enhancing image realism significantly. My expertise in TensorFlow and Python, coupled with a strong foundation in teamwork and problem-solving, led to a notable reduction in manual inspection time. Leveraging deep learning and neural networks, I drive innovations that align with employer-sought skills.

Machine Learning Specialization (DeepLearning.AI and Stanford University)

Sentiment Analysis using Transformer

September 2024- October 2024

OutperformAI- Vadodara, Gujarat

● Architected a cutting-edge sentiment analyzer leveraging Transformer architecture, expertly calibrating hyperparameters to unlock 90% accuracy in sentiment classification, a 20% improvement over state-of-the-art benchmarks.

● Crafted a scalable sentiment analysis pipeline, seamlessly integrating pre-trained language models with custom features, to deliver a 30% increase in F1-score and a 25% reduction in inference time.

Image Classification using ConvNeXt

July 2023- October 2023

Parul University- Vadodara, Gujarat

● Engineered a ConvNeXt architecture and designed a data pipeline to train it on CIFAR-10 dataset, achieving an impressive 91% accuracy on the test set, surpassing baseline models by 27%.

● Applied dropout layers, effectively reducing overfitting and boosting model robustness, resulting in a 27% increment in validation accuracy.