Design, build, and productionize agentic systems (multi-step reasoning, tool orchestration, guardrails) for materials science search, Q&A and information extraction.
Develop, integrate and maintain memory systems, MCP servers and agent skills in a multi-agent environment.
Build evaluation frameworks with domain experts to measure answer quality, extraction accuracy, and retrieval performance.
...
Act as a technical expert in Agentic RAG, Agents & Agent Orchestration, Agent Harness Engineering and driving the exploration and adoption of emerging AI capabilities.
Prototype and develop GenAI-powered solutions across web, desktop, and digital workplace platforms, leveraging built-in AI capabilities and modern agent frameworks to accelerate business value realization.
Design, implement, and optimize agent ecosystems, including agent workflows, multi-agent collaboration, tool integration, memory management, retrieval systems, and orchestration patterns to support complex enterprise use cases.
...
Design, build, deploy, and maintain task‑specific Copilot Studio agents, Power Automate workflows, Power Apps, and Azure AI agent solutions, with increasing emphasis on agent‑first and headless designs where appropriate.
Ensure solutions integrate securely with enterprise data sources, Microsoft 365 services, and downstream systems.
Remain accountable for solution quality, including reliability, performance, security, and long‑term maintainability, across all assets built or supported by the MAPP CoE.
...
Research on LLM Agent methodology, how to build better performance/low cost/stability agents;
Explore cutting-edge technologies related to LLM Agent engineering platform(AgentOps), covering areas such as RAG, Reasoning, Planning, Toolkits, Multi-Agents Coordination.
Construct a high-performance, cost-effective large-model agent service architecture that ensures high service availability.;
...
Design, develop, and deploy AI agent systems capable of task planning, tool usage, and autonomous workflow execution.
Build and optimize RAG pipelines, including document chunking, embeddings, vector search, and retrieval orchestration.
Fine-tune large language models (LLMs) using instruction tuning, supervised fine-tuning (SFT), or reinforcement learning from human feedback (RLHF).
...