- Design and implement the technical architecture for Sentia’s core AI agents, including planning modules, memory/retrieval systems, and tool-use orchestration.
- Apply advanced Reinforcement Learning (RL) or control theory methods to develop agents for adaptive intervention strategies.
- Select, fine-tune, and deploy Large Language Models (LLMs) to power agent reasoning and empathetic communication.
- Develop models for human-agent interaction incorporating principles from cognitive science and behavioral economics.
- Construct simulation environments for pre-training and testing agent policies to ensure system safety.
- Optimize deployment environments for low-latency decision-making and high system resilience.
- Design comprehensive evaluation pipelines including trace-level analysis, offline evaluation, real-time production monitoring, and guardrails.
AWSDockerPython+5 more