Senior AI Engineer

Role Summary

We are looking for a Senior AI Engineer to design, build, and operate production-grade Agentic AI systems that power intelligent assistants and autonomous workflows across regulated enterprise environments.

This role will focus on LLM-based agents, multi-agent orchestration, guardrails, and end-to-end deployment pipelines, with direct exposure to Copilot Studio, LangChain, LangGraph, and enterprise CI/CD practices.

The ideal candidate combines strong Python engineering, agent architecture expertise, and rigorous testing/observability discipline to ensure safe, reliable, and accurate AI responses at scale.

Key Responsibilities

Agentic AI & LLM Engineering

• Design and implement Agentic AI systems using LangChain and LangGraph, including planner executor, router, and evaluator patterns.
• Build multi-agent workflows for intent classification, probing, reasoning, and decision orchestration.
• Develop tool-using agents (retrieval, rules engines, APIs, enterprise services).
• Optimize prompt strategies, state management, memory, and reasoning flows to minimize hallucinations and maximize accuracy.

Copilot & Agent Platforms

• Build and extend agents using Copilot Studio and Power Platform based agent frameworks.
• Integrate custom Python-based agents with Copilot runtime, connectors, and enterprise data sources.
• Collaborate with product and platform teams to operationalize agents across real business workflows.

Guardrails, Safety & Compliance

• Implement AI guardrails including:
  • Policy enforcement
  • Output validation
  • Grounding checks (RAG / knowledge-based verification)
  • Human-in-the-loop and escalation patterns
• Ensure agents comply with enterprise risk, regulatory, and data security standards.
• Design architectures that are auditable, observable, and deterministic where required.

Testing, Evaluation & Quality Assurance

• Build automated agent testing frameworks to validate:
  • Correct intent classification
  • Accurate probing behavior
  • Expected response generation
  • Regression prevention across prompt and model changes
• Implement offline and online evaluation (golden datasets, synthetic tests, confidence scoring).
• Partner with QA and Ops to monitor accuracy, failure modes, and drift.

CI/CD & Platform Engineering

• Develop CI/CD pipelines for AI agents (prompt versioning, agent configs, model updates).
• Support containerized deployments and environment promotion (dev test prod).
• Integrate logging, observability, alerts, and performance metrics for agent behavior.

Required Qualifications

Technical Skills (Must Have)

• Strong Python engineering experience (async, APIs, services).
• Hands-on experience with LangChain and/or LangGraph in real-world agent implementations.
• Experience building Agentic AI systems (not just prompts or chatbots).
• Understanding of LLM tooling, RAG, function/tool calling, and orchestration patterns.
• Experience implementing CI/CD pipelines for ML or AI-driven systems.
• Proven experience in testing LLM outputs and agent behavior.

Platform & Architecture

• Experience with enterprise AI platforms (Copilot Studio, Power Platform, or equivalent).
• Familiarity with microservices, APIs, event-driven systems, and cloud-native design.
• Experience designing governed, production-ready AI architectures.

Preferred / Nice To Have

• Experience with Copilot Studio custom agents or connectors.
• Knowledge of LLMOps / AI Ops practices.
• Experience in regulated domains (financial services, healthcare, compliance-heavy environments).
• Familiarity with evaluation frameworks, agent observability tools, and policy engines.
• Exposure to graph-based reasoning or knowledge graphs.

What Success Looks Like

• Agents consistently generate accurate, policy-compliant, and explainable responses.
• New agent capabilities move from prototype to production safely and quickly.
• CI/CD pipelines catch regressions before agents reach users.
• Guardrails prevent hallucinations and incorrect guidance at scale.
• AI systems are trusted by both business users and risk/compliance teams.