AI Chatbot & Agent Assistants
Langchain-based RAG and agent architectures for production assistants (DataBoss, 2024)
Worked at DataBoss Inc. under Prof. S. Serdar Kozat on AI-powered chatbots and agent assistants for industrial applications, using Langchain.
Highlights
- Designed and shipped AI-driven chatbots and agent assistants for varied production purposes.
- Developed multiple Retrieval-Augmented Generation (RAG) variants and agent architectures, including a self-multi-RAG built on top of self-RAG to improve information retrieval and response quality.
- Explored agent-orchestration patterns balancing latency, factuality, and cost.
This experience grounded my research in the practical realities of LLM deployment — what breaks at scale, which ablations matter, and where evaluation is hardest.