Job Title: AI Engineer / Senior Python Developer – Google ADK (Agent Development Kit)
- Client Location: Philadelphia, PA
- Work Model: Remote
- Duration: 8+ months
- USC GC only
Role Overview
- Design and build production-grade AI agents using Google Agent Development Kit (ADK).
- Develop scalable, multi-agent systems that integrate with internal APIs, data platforms, and third-party tools.
- Deploy and manage agents on Vertex AI Agent Engine with strong observability, evaluation, and reliability practices.
Key Responsibilities
- Design and implement single and multi-agent systems using ADK workflow agents (Sequential, Parallel, Loop)
- Build tool-enabled agents using custom Python functions, OpenAPI/MCP tools, and third-party libraries
- Develop evaluation suites to measure response quality, execution traces, and regression behavior
- Create local developer workflows using ADK CLI and Dev UI for debugging and rapid iteration
- Package and deploy agents to Vertex AI Agent Engine, managing versions, configs, and rollbacks
- Integrate multiple LLMs (Gemini via Vertex/AI Studio; alternatives via LiteLLM, Ollama, or vLLM)
- Implement safety and guardrails (HITL approvals, least-privilege access, input/output validation)
- Collaborate with product, ML, platform, and security teams to define interfaces, SLAs, and success metrics
Required Skills & Experience
- 5+ years of professional Python development experience
- Strong experience building backend services, SDKs, or automation
- Hands-on experience with async I/O and concurrency (asyncio, multiprocessing)
- Experience with API design (REST and/or gRPC)
- Containerization experience with Docker
- Practical experience integrating LLM SDKs, function/tool calling, or agentic workflows
- Strong software engineering fundamentals: observability, retries, idempotency, tracing, performance tuning
- Comfortable working in Linux and cloud environments with secure secrets/config management
Preferred / Nice to Have
- Hands-on experience with Google ADK (google-adk)
- Experience deploying agents to Vertex AI Agent Engine
- Experience configuring Gemini and alternate models via LiteLLM, Ollama, or vLLM
- Familiarity with MCP tools, Agent-to-Agent (A2A) communication, or streaming interactions
- Experience with RAG pipelines, vector stores, and LLM cost/performance optimization
- Prior experience with LangChain, LangGraph, or CrewAI
Tech Stack
- Python: google-adk, FastAPI, Pydantic, Pytest, Ruff/Black, MyPy
- Agents: ADK CLI & Dev UI, workflow agents, tool integrations, evaluation harnesses
- Cloud & DevOps: Vertex AI Agent Engine, Cloud Run, Docker, GitHub Actions
- Models: Gemini (Vertex AI / AI Studio), LiteLLM, Ollama, vLLM
Thanks & Regards,
SAM S
Email: – sam.s@anviktek.com // sam.ss@anviktek.com
|
