AI Engineer
loadshare
Job Description
- AI Agent Development: Design, build, and deploy AI agents that can autonomously query databases, analyse logistics data, perform root cause analysis, and answer business questions for operations, product, and leadership teams.
- MCP Infrastructure: Own and evolve the LoadShare MCP – connecting LLMs (Claude, Gemini, etc.) to Redshift, MySQL read replicas, and internal APIs with proper access controls, read-only security patterns, and schema context.
- Agent Architecture & Evaluation: Evaluate multiple agent architectures on output accuracy, latency, and token cost. Drive decisions on the right patterns for different use cases.
- Workflow Tuning: Optimise agent tool-calling workflows for efficiency – reducing unnecessary LLM calls, improving context retrieval, and ensuring agents produce reliable, validated outputs.
- Knowledge Layer: Build and maintain the shared context layer (table schemas, business rules, metric definitions, join patterns) that powers all AI tools across the org, creating a single source of truth.
- Voice & Conversational AI: Contribute to AI-powered voice agents for grievance classification, rider support, and customer interaction using models like Gemini Live and Claude.
- Internal Tooling & Enablement: Support the broader engineering team in adopting AI tools – building skills, MCP connectors, and reusable skills that make everyone more productive.
- Security & Access Control: Implement robust security patterns for AI-to-database access, including read-only user provisioning, IAM role mapping, DDL restrictions, and query guardrails.