Description
Led a diverse team of 8 including Machine Learning Engineers, Data Engineers and Data Annotators shaping the company’s AI product offerings and driving feature enhancements.
Automatic Essay Feedback : Fine-tuned a pre-trained T5 and GPT-2 model for automatic essay feedback generation using a curated dataset of essay feedback tailored for college admission essays improvement, resulting in 20% faster turnaround time and 37% reduction in editing service cost to company
Trained Adapter Transformers (Adapter Fusion, Parallel Adapter, Mix-n-match Adapters) for low-resource model training and tackling data-scarcity by leveraging learning from data-rich classes. Effectively reduced deployed model parameters by one-eighth.
Lead author for filed patent: MAURYA, Himanshu, et al. Machine Learning System for Analyzing the Quality and Efficacy of Essays for Higher Education Admissions
Awarded “A Beautiful Mind” award for Commitment to Innovation for H2 2020 by iSchoolConnect, April 2020