AI Engineer – Agentic Systems & RLAIF

Capgemini is hiring an AI Engineer – Agentic Systems & RLAIF for an onsite contract role in Dallas, TX. In this position, you will help build and operate an RLAIF plane that turns production evidence from customer-facing agents into safer, measurable, and reversible improvements. The work emphasizes governed agent runtime, strong observability, and evaluation harnesses, deployed on Google Cloud with calibrated AI judging and root-cause-driven optimization. Salary is USD 39–61 per hour.

What you’ll do

• Build and run production agents: design and operate production agents and multi-agent workflows using LangGraph, LangChain, and deep-agent patterns, with governance, monitoring, and versioning from day one.
• Instrument for observability: implement tracing, telemetry, and failure taxonomies across agents, tools, memory, and knowledge sources so failures can be traced back to root causes.
• Own evaluation harnesses: manage golden sets and drive both offline and online evaluations, including AI-judge pipelines with order-swap bias control and human calibration loops, with reporting focused on business outcomes.
• Optimize through governed experimentation: use observed failures to improve prompts, tools, skills, and rules via one-change-at-a-time experiments with regression-safe releases, including hard gates, approval, canary, and rollback.
• Separate optimization paths: keep harness, contextual-policy, and model-level changes distinct and independently evaluated, with mechanisms to detect capability and knowledge-coverage gaps and recommend governed updates.
• Ship on GCP: deploy and scale the RLAIF plane on Vertex AI, GKE or Cloud Run, BigQuery, and Pub/Sub with CI/CD, Infrastructure as Code, and cost/latency controls.
• Collaborate and lead: partner with architects, data scientists, SRE, and client business stakeholders; document decisions, mentor engineers, and raise engineering practice for agent systems.

In addition, you will support root-cause investigation and improvement workflows, including ingesting and analyzing traces, reconstructing the sequence of events, tool calls, and business flow, and driving prompt/context/routing variants backed by evaluation dimensions and calibrated AI judges. You will also define scoring and reward-signal frameworks and generate autonomous optimization recommendations.

Root-cause themes you may address

• Hallucinations, context drift, poor prompts, misrouting
• Policy and compliance issues; safety topics such as harmlessness, different biases, and toxicity
• Missing or underperforming tools
• Taxonomy or ontology gaps
• Re-configuration opportunities for identification and recommendation to improve project outcomes

Requirements

• Demonstrable experience building and running AI agents in production, including shipped systems with ownership of reliability, latency, cost, and incident response.
• Hands-on experience with AI observability and evals, including tracing frameworks (for example, LangSmith, Arize Phoenix, OpenTelemetry), LLM-as-judge design, golden-set curation, and regression testing.
• Deep practical experience with LangGraph and LangChain, including deep-agent patterns (planning, sub-agents, tool use, memory, human-in-the-loop, and state persistence).
• Strong Python engineering skills, including async services, APIs, containers, and disciplined testing in a CI/CD environment.
• Working understanding of RLAIF/RLHF concepts, prompt and policy optimization, and isolating a single causal change in a complex system.
• 5+ years software engineering experience, including 2+ years focused on LLM or agentic applications.

Technologies

Python, LangGraph, LangChain, LangSmith, Arize Phoenix, OpenTelemetry, Google Cloud, Vertex AI, GKE, Cloud Run, BigQuery, Pub/Sub, CI/CD, IaC, LLM-as-judge

Preferred qualifications

• Cloud experience on Google Cloud (Vertex AI, GKE, Cloud Run, BigQuery, Pub/Sub, IAM); GCP certification is a plus.
• Experience with retrieval, enterprise knowledge bases, and memory systems with provenance, trust scoring, and decay.
• Familiarity with governed release practices in regulated or large-enterprise settings, including approval workflows, canary, rollback, and audit trails.
• Telecom or large-scale customer-service domain exposure, plus experience with MLOps tooling and feature flags.

Benefits

• Medical, dental, vision, and retirement benefits