Mahikshit Kurapati

Experience

• 

  Defence Research and Development Organisation (DRDO)

  Dec 2021 - May 2022

  Machine Learning Intern

  • Spearheaded mission-critical audio interpretation by developing an advanced Audio Captioning model using LSTM and Librosa, achieving 89% accuracy in high-noise environments and significantly boosting situational awareness capabilities for critical operations.• Elevated drone reconnaissance capabilities by engineering a Deep Learning model for heat signature analysis, achieving 3x speeds in object and personnel identification across challenging terrains, which boosted operational effectiveness by 70% in complex urban environments. Show less

• 

  Intripid (formerly Travana)

  Sept 2023 - Oct 2024

  Machine Learning Engineer

  Key accomplishments include:* Architected a multi-agent RAG model that integrated diverse data sources such as YouTube and TripAdvisor, resulting in a remarkable 50% increase in user satisfaction alongside a substantial 35% rise in itinerary completion rates. * Implemented advanced NLP techniques using BERT for vector embeddings that refined information retrieval processes by an impressive 40%, cutting average query response times from five seconds to three seconds effectively. * Optimized data fetching through asynchronous programming strategies that decreased static data retrieval time by approximately 30%, enabling a concurrent user capacity increase of about 25% without additional hardware costs. * Developed a regression modeling-based predictor achieving an outstanding accuracy rate of 91% for global destination recommendations, driving a significant spike of around 60% in user engagement with suggested locations. * Designed an Information Retrieval Algorithm for the RAG model that enhanced query relevance metrics by roughly 25%, while simultaneously reducing response times by around 20%, which led to a notable decrease of about 40% in irrelevant recommendation reports from users.Skills: Machine Learning, Retrieval-Augmented Generation (RAG), Natural Language Processing (NLP), Asynchronous Programming Show less

• 

  Consolidated Communications

  Oct 2024 - now

  AI/ML Engineer

  Key Responsibilities:Design and Deploy AI-Powered Solutions:Lead the development of multi-agent virtual assistants using Python, Langchain, and OpenAI APIs to significantly reduce average customer handling time and enhance satisfaction. Integrate legacy CRM systems to create seamless, intelligent customer support workflows.Develop Advanced Notification Systems:Utilize OpenAI GPT-4o API, Twilio, and PEGA to build SMS notification and customer management systems with agentic workflows. Implement techniques like Chain-of-Thought (CoT) and CoT+SC tuning to minimize hallucinations and ensure accurate responses.Automate Reporting and Analytics:Automate the generation of weekly business reports using PostgreSQL queries in pgAdmin, ensuring data accuracy and reducing errors in LLM-user interactions by 40%.Build Intent Classification Models:Develop high-accuracy intent classification models using BERT and advanced NLP techniques such as tokenization and domain-specific fine-tuning. Achieve 96% accuracy in classifying user intents (e.g., update order, cancel, postpone, create).Foster Team Knowledge and Adoption:Conduct knowledge transfer (KT) sessions to educate the team on effectively utilizing LLMs and understanding their value in enhancing customer support processes.Technical Skills:Programming Languages: Python, SQL (PostgreSQL)AI/ML Frameworks: Langchain, OpenAI API, BERT, TensorFlow/KerasNatural Language Processing (NLP): Tokenization, Fine-tuning, Intent ClassificationAutomation Tools: pgAdmin, Twilio, PEGAAPI Integration: OpenAI GPT-4o, Twilio Show less