Harshit Gupta

Data Scientist with 3.5+ years of experience, highly motivated and results-oriented. Skilled in applying machine learning, deep learning, and natural language processing techniques to solve complex problems and improve patient treatment outcomes. Pioneered a real-time clinical NER automation POC, slashing turnaround times and expediting urgent customer actions. Revolutionized patient treatment through the development of a novel, patent technology that leverages predictive biomarkers for precision medicine. Experienced in Agile Development and proficient in Python, R, SQL, C++, and Linux. Strong background in databases such as MongoDB, ElasticSearch, Hive, Snowflake, and Azure File Storage. Passionate about LLMs, Retrieval Augment Generation (RAG) Pipelines, Vetor Databases, Natural Language Processing, Traditional ML Algorithms, Graph ML, Sequence-to-sequence Modelling, LLM Finetuning, and Agile Development.

At Optum Global Solution, I designed and implemented various AI-powered solutions, including an Advocate performance prediction POC and an AI-powered phrase-based search tool. I also achieved a Rouge score of 0.78 for a summarization model on Conversational data and developed a keyword search algorithm leveraging Pyspark. As a Research Analyst at INNOPLEXUS Consulting Pvt. Ltd, I developed and deployed a high-accuracy Biomarker prediction model and automated literature data mining for biomarker extraction. I also achieved 93% test accuracy on a Bert based Transformer model for chemical smile activity prediction. As an AI Developer Intern at E-Unimart Multichannel Pvt. Ltd, I achieved 85% accuracy on a product similarity model and coordinated a team to develop a customized word embedding e-commerce description data.