Senior Applied AI Engineer

Role Overview

Build production AI systems for clients. You'll design and ship agentic workflows on Claude and AWS Bedrock AgentCore, retrieval systems backed by Bedrock Knowledge Bases, and MCP-based automations that connect directly into client tool stacks (Notion, Linear, HubSpot, AWS, Google Workspace, and more). This is a hands-on build role. You will write code, deploy infrastructure, and sit in client rooms explaining why the system behaves the way it does.

Tasks & Responsibilities

• Design and build applied AI systems on Claude and AWS Bedrock AgentCore: RAG pipelines backed by Bedrock Knowledge Bases, agentic workflows, and multi-tool orchestration via MCP.
• Own delivery of AI components end-to-end: architecture, implementation, evaluation, and post-launch monitoring.
• Integrate Claude-based agents with client infrastructure and business tools via MCP servers (HubSpot, Notion, Linear, Google Workspace).
• Translate ambiguous client problems into scoped technical builds with clear success criteria and numbers attached.
• Document architecture and decisions in the client's Notion record so infra details never live only in someone's head.
• Work alongside Pre-Sales Engineers during scoping calls to validate technical feasibility before commitments are made.
• Select and tune the right model, retrieval strategy, and Knowledge Base configuration per use case.
• Build and maintain eval suites (offline test sets, regression checks, LLM-as-judge scoring) that catch quality drift before a client does.
• Design agentic control flow: tool-selection logic, retry/fallback paths, and guardrails against runaway loops or unsafe tool calls.
• Instrument production systems with logging, tracing, and cost/latency dashboards to debug a bad output from a client.
• Write modular infrastructure as code (Terraform) for AI workloads on AWS.
• Conduct architecture reviews on other engineers' AI builds and flag failure modes before they ship.

Competencies

• Can explain to a client, in plain language, why an agent made the tool call it made
• Comfortable debugging a production LLM failure under time pressure, not just in a notebook
• Can wire an MCP server into an agent and debug why a tool call didn't fire

Education, Experience & Language

• Master's degree in Computer Science or equivalent
• Previous experience building and shipping production ML/AI systems
• Fluent in English and Arabic

Tools & Certifications

• Strong Python engineering; comfort with AWS
• Practical experience with LLM application patterns: RAG, agentic tool-use via MCP, prompt engineering, evaluation design
• Hands-on experience with AWS Bedrock AgentCore and Bedrock Knowledge Bases
• Nice to have: AWS Certified Machine Learning Engineer - Associate, AWS Certified Generative AI Developer - Professional, AWS Certified Solutions Architect - Associate or Professional, AWS Certified Developer - Associate, AWS Certified Security - Specialty (relevant given our compliance posture)