aman varyani

I am a Generative AI engineer with hands-on experience building production RAG pipelines, vector-database retrieval, and multi-agent chatbots. I design systems that combine LLMs, vector stores, and cloud tools to deliver reliable user-facing AI products.

At my current role I implemented a full RAG pipeline using pgvector, built an agentic chatbot capable of web searches, and created prospect and email-editing bots within a multiagent architecture. I also applied RAGAS and LangGraph evaluations and worked with fine-tuning and RLHF methods.

Previously, I integrated HubSpot and Salesforce, developed Docker services for document thumbnails, and optimized Django applications with caching and task queues to improve load and response times. I created reusable class-based views, models, and serializers and implemented JWT authentication to streamline user login.

I led an open-source AI career assistant project that earned 350+ GitHub stars, demonstrating strong community engagement and practical impact through automated coverletter customization and company research via a Streamlit interface.