AI Engineer III
Mariner Finance · Plano, TX, US
About this role
**AI Engineer III — Mariner Finance** **Base Salary Range:** **$133,836.99 – $150,000.00 / Yr** --- ## About the Role In this role, you will build and maintain the **knowledge layer** and **backend services** that power **enterprise AI agents**—turning standalone models into **reliable, governed, tool-using agents**. --- ## Key Responsibilities - Build and maintain enterprise **knowledge layer** components for **grounding, memory, and context** (skills, knowledge graph elements, vector indexes). - Develop and maintain the **skills and tool registry** (definitions, versioning, discovery, publishing workflows) and build **MCP servers/APIs** to expose capabilities. - Implement **knowledge graph pipelines** (entity extraction, entity resolution, relationship loading, schema updates) and build **graph/hybrid retrieval** queries. - Build and tune **vector search** (embeddings, semantic chunking, metadata filtering, hybrid search, index refresh) and measure retrieval quality against agreed metrics. - Create **batch + streaming data pipelines** on **Databricks** (Delta, Spark, Structured Streaming) and modern ETL/ELT patterns to ingest structured, semi-structured, and unstructured data. - Integrate business systems and document stores into the knowledge layer via **secure REST APIs/connectors**. - Register and govern data/AI assets using **Unity Catalog**; implement **access controls/entitlement checks** so agents only return permitted data. - Build **agent evaluation suites** (golden datasets, synthetic tests, LLM-as-judge, rule-based scorers, regression tests) covering retrieval quality, groundedness, tool reliability, safety, latency, and cost. - Integrate evaluations into **CI/CD** as quality gates; review results to identify regressions and improvements. - Write **production-ready Python** and robust web APIs (**Flask/FastAPI**) for agent actions and knowledge services. - Implement agentic tools/workflows for **Databricks Custom Agents**; apply context/prompt engineering for context windows and structured outputs (e.g., JSON schemas). - Add **monitoring, tracing, and logging** for agent loops, tool usage, and retrieval systems; follow security practices (identity, secrets, runtime isolation, data protection). - Support **CI/CD**, testing triggers, multi-environment deployments; deploy on **AWS** (including **EC2**) and use infrastructure-as-code where available. - Maintain technical documentation, runbooks, and lineage records to support auditability and responsible AI governance. - Troubleshoot issues across pipelines, retrieval quality, integrations, and agent behavior; escalate complex problems as needed. --- ## Required Qualifications - Bachelor’s degree in CS/Engineering/Math/Stats/IS or related field (or equivalent experience). - **6+ years** in software engineering, data engineering, platform engineering, ML engineering, or related—hands-on AI/ML/LLM experience. - Strong **Python** (async + testing) and experience building **REST APIs** (Flask/FastAPI). - Experience with **ETL/ELT + SQL** (preferably Databricks or similar like Snowflake). - Working knowledge of **vector search/RAG** (embeddings, chunking, retrieval pipelines). - Knowledge of **knowledge graphs/graph databases** (e.g., Neo4j, Amazon Neptune, RDF/SPARQL) and graph/ontology modeling. - Experience building **evaluation datasets** and automated scoring for LLM/agent outputs. - Familiarity with **agentic tooling** (tool calling, structured outputs, agent frameworks
Listing freshness
CronJobs last confirmed this listing 1h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.