Akshay Kalra

Developing an advanced resume management system using retrieval augmented generation, focusing on fine-tuning Computer Vision and NLP models. Trained statistical models (XGBoost, generalized linear models) for predictive analytics on customer engagement. Executed CV techniques including Image Processing, Image Classification (CNN), Object Detection (Faster R-CNN), and OCR Engine for digitizing information from Floor Plan Images. Developed a Text Classifier using a Roberta-Base model (F1-score of 0.95) and collaborated on fine-tuning Deep Attractiveness Rank Net (DARN) and engineering a personalized tagging tool.

Contributed to cost reduction (INR 140 million) by developing automation solutions using Classical ML algorithms. Automated L1 & L2 support engineer workflows by deploying a support ticket classifier tool across 20 projects, reducing time spent on ticket routing by 30%. Implemented various classifiers (Logistic Regression, Support Vector Classifier, Random Forest Classifier, XGBoost) and designed a Text Similarity algorithm using Smooth Inverse Frequency, Glove Word Embeddings, and Tf-Idf Vectorization. Developed a Named Entity Recognition Model to identify servers, individuals, and application names within service ticket descriptions.