Description

Design and build specialized LLM agent systems for complex data retrieval, reasoning, and validation.

This role is on-site.

Responsibilities

• Design and implement Multi-Agent Systems (MAS) using Google Agent Development Kit (ADK) for multi-step reasoning tasks.
• Architect Human-in-the-Loop (HITL) workflows, managing hand-offs between agents and human experts.
• Implement advanced reasoning patterns including Chain-of-Thought (CoT), ReAct, and Self-Reflection to improve output transparency.
• Optimize long-context window utilization for agent coherence when analyzing vast datasets.
• Establish rigorous AI Evaluation (Eval) frameworks to measure agent performance via trajectory analysis and faithfulness.

Required Skills

• 6+ years of experience in AI development, preferably on Google Cloud Platform.
• Strong proficiency in Python programming.
• Hands-on experience building agents using Google ADK or similar agentic frameworks.
• Experience with LLM tuning techniques such as SFT, CoT, and ReAct.
• Proven track record implementing AI evaluation methods and metrics (Precision, Recall, F1 score for trajectories).
• Experience managing sequential and loop-based agentic workflows.
• Ability to debug and perform Root Cause Analysis (RCA) for agentic issues.
• Experience deploying AI solutions within an agent-based framework.
• Familiarity with Agile development processes.

Preferred Skills

• Experience with CI/CD pipelines for AI model deployment.

Key Skills

Education

Any Gradute