About this role
Design, build, and evaluate agentic AI systems (single- and multi-agent) that plan, reason, act, and collaborate across tools and environments, and develop robust agent harnesses and evaluation frameworks. Implement self-improving agent loops, architect agent memory systems, and enable reliable deployment of agents on constrained and edge environments through model/runtime optimization and secure tool-use.
Skills for this role
PythonLarge language models (LLMs)Agentic systemsTool-using agentsPlanner-executor patternsMultistep reasoning pipelinesRetrieval-augmented generation (RAG)PromptingModel evaluationError analysisSoftware engineeringExperiment designAblation studiesRegression analysisTrace loggingReplayabilitySuccess and performance metricsSelf-improving agentsReflectionCritiqueSelf-debuggingAgent memory systemsSummarizationCompressionPrivacy-aware retentionModel/runtime optimizationEdge deploymentOffline executionSecure tool useEdge-cloud coordination (edge-cloud coordination is mentioned in source textre