Artificial Intelligence Engineer

Insight Global; Atlanta (GA); On-site; USD 120,000 - 180,000; Full time

Job summary

Insight Global is seeking an experienced software engineer to work across the AI layer of our platform, contributing to retrieval pipelines, agent workflows, and model evaluation. The role is hands‑on and empirical, with experimentation, measurement, and iteration driving progress, in collaboration with the Senior AI Engineer and broader engineering team.

You will implement and tune retrieval components, build out the agent orchestration layer, and instrument observability across AI systems.

Qualifications

• 3+ years of software engineering experience, with at least 1 year building LLM-powered systems.
• Practical experience in retrieval-augmented generation or agent workflows.
• Proficient in Python with asynchronous programming and REST APIs.
• Familiarity with evaluation of LLM systems and observability tooling.
• Knowledge of RAG pipelines, agent orchestration, or fine-tuning.

Responsibilities

• Contribute to the hybrid retrieval pipeline — implementing and tuning retrieval components, running experiments to improve quality, and validating results on real evaluations.
• Build and maintain components of the agent orchestration layer — tool integrations, prompt management, and supporting human‑in‑the‑loop workflows.
• Instrument AI and agent systems for observability — tracing model and tool calls, capturing token and latency telemetry, and supporting failure analysis.
• Build and maintain evaluation datasets and test suites across retrieval and agent workflows, contributing to CI‑level quality gates.
• Support document AI capabilities — working with parsing, extraction, and OCR pipelines as the platform serves new client engagements.
• Implement per‑tenant isolation checks across retrieval and agent layers, helping ensure no cross‑tenant data leakage occurs.
• Contribute to model behavior evaluation — running prompt injection and adversarial tests as part of ongoing eval work.

Skills

Python

Async code

REST APIs

LLM systems

RAG pipelines

Tools

LangGraph

OCR pipelines

Evaluation datasets

Job description

About the role You will work across the AI layer of the platform — contributing to retrieval pipelines, agent workflows, and model evaluation. The work is hands‑on and empirical: you run experiments, measure results, and iterate. You will work closely with the Senior AI Engineer and broader engineering team, taking increasing ownership as you develop depth across the platform's AI systems.

Key Responsibilities:

• Contribute to the hybrid retrieval pipeline — implementing and tuning retrieval components, running experiments to improve quality, and validating results on real evaluations.
• Build and maintain components of the agent orchestration layer — tool integrations, prompt management, and supporting human‑in‑the‑loop workflows.
• Instrument AI and agent systems for observability — tracing model and tool calls, capturing token and latency telemetry, and supporting failure analysis.
• Build and maintain evaluation datasets and test suites across retrieval and agent workflows, contributing to CI‑level quality gates.
• Support document AI capabilities — working with parsing, extraction, and OCR pipelines as the platform serves new client engagements.
• Implement per‑tenant isolation checks across retrieval and agent layers, helping ensure no cross‑tenant data leakage occurs.
• Contribute to model behavior evaluation — running prompt injection and adversarial tests as part of ongoing eval work.

Required Qualifications:

• 3+ years of software engineering experience, with at least 1 year building LLM‑powered systems — RAG pipelines, agent workflows, or fine‑tuning — in a production or near production setting.
• Practical experience in at least one of: retrieval‑augmented generation (RAG) and reranking; agent orchestration with LangGraph or comparable; or LLM fine‑tuning.
• Proficient in Python and comfortable working with async code, data pipelines, and REST APIs.
• Exposure to evaluation methodology for LLM systems — has contributed to or authored an eval dataset or test suite.
• Familiarity with agent or LLM observability tooling — tracing, logging, or monitoring of model and tool calls.
• Working knowledge of modern LLM and information‑retrieval concepts; can discuss trade‑offs between approaches with evidence.

Preferred Qualifications:

• Experience with knowledge graphs or property‑graph query languages.
• Exposure to document AI — PDF parsing, table extraction, or OCR pipelines.
• Has contributed to or built an evaluation dataset and labeling workflow.
• Familiarity with prompt‑injection risks and mitigation strategies in agent and retrieval pipelines.
• Open‑source contributions to LangGraph, sentence‑transformers, or comparable projects.