Aviral Sharma is a data science and machine learning professional with over 5 years of experience developing large-scale, end-to-end solutions across healthcare, finance, and telecom. He has a proven ability to deploy containerized applications using Kubernetes and has developed multiple novel information extraction solutions, resulting in filed patents. His expertise includes advanced NLP, deep learning models, and optimizing complex business processes.
Experience
Data Scientist
Optum
Devised language models utilizing NLP meta features to extract key clauses from scanned contracts, and enhanced the solution using a custom NER model. Deployed the end-to-end solution on Kubernetes for processing over 200k contracts. Also productionized a search engine for 21k PBM contracts on Kubernetes, utilizing MongoDB indexing and Elasticsearch. Developed a vocabulary agnostic information extraction and linking engine.
Analytics Consultant
FN MathLogic
Worked with a major Australian telecom provider to develop a national field technician productivity enhancement model using XGBoost and Bayesian Optimization. Developed Kubernetes-based Kubeflow Pipelines and utilized a GIT-lab based CICD pipeline to deploy the model in production, achieving a 5% productivity enhancement. Other roles included implementing ML risk models for credit default assessment and devising claim denial management models.
Education
Indian Institute of Technology Delhi
B. Tech
Textile Engineering