Backend AI Engineer (W2 Only)

Job Title - Backend AI Engineer
Location - 100% Remote PST
Duration - 12 Months

Job Description
Looking for Backend Engineer with strong core Python and SQL skills who can create end to end applications, along with solid fundamentals in Agentic AI.

• Data structures.
• End to end API development
• CURD operation
• Session management
• Context management
• Sync, Async, callbacks, Streaming
• SQL, Queries, Nested Queries, Joins
• Experience on Data cleaning and pipeline creation.
• Solid understanding in Agent and Multi-agent development.
• Knowledge of framework and Protocols like A2A, MCP, ADK etc.

Job Responsibilities
Al Systems Engineer - Architects and engineers the "intelligence engine" by building autonomous agentic frameworks, high-scale data pipelines, and the cloud infrastructure required for production. Bridges the gap between raw data and agentic action, ensuring the system is secure, scalable, and seamlessly integrated into the enterprise ecosystem. [Needs to be experienced in designing scalable solutions on GCP and extra points if has G3 Experience.
Agentic Logic & Orchestration: Designs and implements multi-step reasoning loops and autonomous tool-calling frameworks (e.g., ADK, MCP) that allow Al to execute complex business tasks and "handshake" with external APIs without manual intervention.
Data Pipeline & Artifact Engineering: Develops production-grade SQL/Python pipelines to ingest, clean, and structure high-velocity data, maintaining the vector databases and data artifacts that ensure the Al is grounded in real-time, accurate information.
Cloud Infrastructure & Deployment: Provisions secure, enterprise-ready cloud environments using Infrastructure as Code (laC) and manages containerized deployments (K8s/Docker) to ensure high availability and low-latency performance at scale.
Al Evaluation & Automated Quality: Implements automated "Eval" frameworks (e.g., Promptfoo, LangSmith) to quantitatively measure model accuracy, grounding, and hallucination rates, ensuring the agent's reasoning is statistically sound and safe for a high-stakes enterprise rollout.