AI Engineer

Job Title: Senior AI Engineer – Agentic Systems & Orchestration
Experience Level: Senior (5+ years of software engineering experience)
Position Overview
We are seeking a AI Engineer to design, build, and maintain production-grade agentic systems and robust AI harnesses. In this role, you will bridge traditional software engineering (Python, .NET, modern frontend) with advanced Generative AI pattern design. You will build deterministic control loops around non-deterministic LLM behaviors, implement end-to-end workflow orchestration, establish full observability, and create automated evaluation pipelines to ensure high-reliability AI deployment at scale.
Key Responsibilities
Agentic Architecture & Harness Engineering: Architect robust agent execution harnesses that manage context, enforce safety boundary limits, implement tool calling, and manage dynamic execution loops (ReAct, reflection, plan-and-execute).
Workflow Orchestration & Loops: Implement complex multi-step workflows using stateful orchestration engines. Design resilient human-in-the-loop (HITL) checkpoints, automated retry logic, backoff mechanisms, and self-correcting execution loops.
Full-Stack Implementation: Develop core microservices and APIs in Python and .NET (C#) while building responsive, real-time user interfaces (React/TypeScript, Vue, or Angular) for stream tracking and human intervention.
Observability & Tracing: Instrument full-stack telemetry across agent execution graphs. Track trace IDs, latency bottlenecks, token consumption patterns, tool invocation success, and intermediate decision steps across complex agent runs.
Evaluation & Testing Frameworks (Evals): Build automated evaluation pipelines using LLM-as-a-Judge, unit evals, regression test suites, and dynamic benchmarks to measure agent accuracy, hallucination rates, and task completion metrics before release.
Required Qualifications
1. Software Engineering Core
5+ years of professional software engineering experience delivering production software.
Backend: Advanced proficiency in Python (FastAPI, Django, Pydantic) and .NET / C# (ASP.NET Core, Task Parallel Library, dependency injection patterns).
Frontend: Strong experience with modern JavaScript/TypeScript frameworks (e.g., React, Angular, or Vue.js) to construct interactive, real-time monitoring dashboards or agent human-in-the-loop interfaces.
Architecture: Firm understanding of asynchronous processing, queue systems (RabbitMQ, Kafka, Azure Service Bus), distributed state management, and REST/WebSocket APIs.
2. Agentic Systems & AI Concepts
Agentic Patterns: Hands-on experience building autonomous or semi-autonomous agentic workflows, function calling pipelines, long/short-term memory structures, and iterative loop controls.
Workflow Orchestration: Experience using frameworks designed for graph/state control and workflow management, such as LangGraph, Temporal, Semantic Kernel Process Framework, AutoGen, or LlamaIndex Workflows.
Observability & Monitoring: Hands-on integration experience with telemetry and tracing platforms like LangSmith, Phoenix (Arize), Langfuse, or OpenTelemetry.
Evaluation & Testing: Proven track record setting up continuous evaluation (Evals) using tools such as Ragas, DeepEval, Promptfoo, or custom LLM-as-a-Judge test runners integrated into CI/CD pipelines.
Preferred Qualifications
Deep experience deploying on cloud platforms (Azure OpenAI, AWS Bedrock, or GCP Vertex AI).
Familiarity with vector databases (Pinecone, Qdrant, PGVector) and hybrid search mechanisms.
Experience with guardrail systems (Guardrails AI, NeMo Guardrails) to constrain agent behavior and enforce runtime safety.
Background in deterministic fallbacks for non-deterministic model outputs.
What We Offer
Competitive salary and performance-based equity/bonus.
Comprehensive health, vision, and dental coverage.
Flexible work setup (Remote / Hybrid).
Annual learning allowance for research, AI tools, and technical conferences.

Work arrangement
Yes