Prashant Shukla

Led LLM deployment on AWS, optimizing scalability for language tasks. Implemented open-source guardrail AI for enhanced content filtering. Fine-tuned "Falcon-7b" LLM into a versatile chatbot, boosting performance. Leveraged langchains to build AI agents, integrating LLM with APIs and OCR. Applied advanced NLP for valuable insights from textual data, improving decision-making and customer communication, specializing in voice data processing and ASR. Implemented robust data pipelines, incorporating statistical techniques and hypothesis testing to enhance model performance and ensure accurate insights for decision-making. Collaborated on cross-functional teams for NLP and voice data projects, designing algorithms for language identification, entity recognition, and question-answering models. Crafted a state-of-the-art conversation management system utilizing slot filling and deep learning, adeptly capturing vital conversational contexts like intent, entities, and dialogue history. Implemented NLP tasks, including language identification, domain/intent classification, entity recognition, question-answering models (Zero Shot BERT, LSTM), and various ML models for document classification, NER, and semantic keyword matching. Collaborated across teams, achieving a 21% reduction in churn and a threefold increase in annual revenue for a key user segment.

Involved in database migration and developed automation pipelines in Python and Shell scripting. As cohort head managed the team and addressed queries.

Developed a recommendation system for an Ed-Tech community, allowing each mentee to receive a curated list of mentors based on information from resumes, video conferences, forms, and LinkedIn profiles.