RLHF & SFT for Customer Service Support Agents (Zendesk & Intercom Integration)
Leveraged my background in MSCS (Georgia Tech) and Customer Experience Management to improve the accuracy of LLM-driven support bots. I specialized in Supervised Fine-Tuning (SFT) and RLHF to ensure that automated responses maintained high empathy while adhering to complex SLA and security protocols. My work involved evaluating AI-generated ticket summaries and 'ranking' responses based on tone, technical accuracy, and fraud prevention measures.