Subash Pandey

Develops and maintains AI receptionist and interviewer platforms, including conversational AI features, question generation, grading logic, prompt engineering, and production bug resolution. Leads development of a Generative AI sales chatbot integrated with Zoho CRM using LangGraph, RAG pipelines, and Qdrant for lead qualification, intent detection, information extraction, lead scoring, and SEO tracking. Builds chatbot and RAG evaluation pipelines that improved response accuracy by 20%. Conducts audio feature extraction research using Parselmouth, Librosa, and Praat. Also contributes to a full-stack SEO SaaS platform by developing backend services, database migrations, billing, authentication, and AI-driven keyword analysis features.

Designed and developed domain-specific chatbots for banking, travel, and customer support. Implemented Retrieval-Augmented Generation and multi-agent frameworks, integrated databases and external knowledge sources, fine-tuned language models for five use cases, and managed chatbot projects from requirements through deployment. Conducted testing and optimization for accuracy and performance and deployed interactive web chat applications for production use.

Engineered applied machine learning solutions for business problems, built and deployed scalable ML models in cloud environments, and collaborated with data science and engineering teams on production-ready solutions. Applied frameworks and tools including HuggingFace, PyTorch, and AWS.