Agentic AI Architect
San Jose, California, United States of America
Contract
4 hours ago
Key skills
PythonJavaAIGeminiAgenticBigQueryGoogle CloudSpannerPub/SubVertex AI
About this role
We are looking for Agentic AI Architect for our client in San Jose, CA.
Job Title: Agentic AI Architect
Job Location: San Jose, CA
Job Type: Contract
Job Overview:
Pay Range: $60/hr - $65/hr
Requirement/Must Have:
- Minimum 10+ years of professional software engineering experience with a proven track record in full-stack or backend development.
- Deep practical knowledge of Java and Python, demonstrated through significant development experience.
- Hands-on experience with LLMs, specifically the Gemini family, and their application in agentic workflows, reasoning tasks, and tool-calling.
- Familiarity with Agent Development Kit (ADK), Model Context Protocol (MCP), and agent design in Vertex AI.
- Strong understanding of cloud-native development on Google Cloud Platform.
- Demonstrated ability to lead complex projects from requirement gathering to production release, acting as a strategic liaison between engineering and product stakeholders.
Responsibilities:
- Architect and orchestrate autonomous agents capable of planning, executing, and verifying complex engineering tasks such as code generation, unit test case creation, and automated bug triaging.
- Develop and maintain Model Context Protocol (MCP) servers to connect agents to canonical internal documentation and tools, ensuring high-fidelity, grounded responses.
- Utilize the Google Antigravity platform and Agent Development Kit (ADK) to build context-aware assistants that integrate with existing IDEs and development tools.
- Lead the transition of AI Proof-of-Concept (POC) projects into production-ready services, collaborating with internal search and productivity teams.
- Propose and implement architectural changes to agentic systems to improve reasoning capabilities, reduce latency, and optimize token usage.
- Guide junior engineers in adopting 'vibe coding' practices and leveraging Gemini-powered agentic tools for effective software development.
Nice to Have:
- Experience in Google-proprietary technologies such as Borg, Blaze, Piper, and Critique.
- Cloud-native development on Google Cloud Platform, including experience with Spanner, BigQuery, and Pub/Sub.