Milpitas, California, United States
• End-to-end development of an internal beta LLM agent, aligning technical design with evolving enterprise product requirements
• Integrated support for multiple LLM backends (OpenAI, Bedrock, Qwen/Ollama), including model configuration and prompt routing logic
• Built internal APIs and endpoints for dynamic query rewriting, memory ingestion, and profile-driven personalization
• Collaborated cross-functionally on feature planning, agent architecture, and debugging runtime performance issues