Yash Singh

Reinforcement Learning for Network Routing: Conceptualized MVP requirements for an AI-defined Network Routing product and performed hyper-parameter tuning on Deep Q Learning Models. Custom Search Engine and QnA Chatbot: Built a search engine for Splunk dashboards and a QnA Chatbot using FastText word embeddings, BM25 algorithm, and HuggingFace sentence transformers. AIOps: Developed a time-series forecasting POC for CPU utilization using an LSTM model with a custom loss function and Streamlit visualization. Food Volume Estimation: Developed a regression model using transfer learning and computer vision concepts like image segmentation and occlusion detection.

Implemented a logistic regression model to forecast no. of covid cases. Created several visualizations for the same using Seaborn, Matplotlib and Tableau.

Saliency Detection Using GANs: Built a model to detect salient components of an image using GANs, implemented SalGAN in TensorFlow 2.0 for distributed training on GCP. Gender Detection Using Transfer Learning: Improvised gender detection service using transfer learning and model retraining on Keras, enhancing accuracy by 8 percent.