Onsite Artificial Intelligence/Machine Learning Engineer Specialist

Build Reusable Foundational Agents: Develop the core data ingestion and processing agents (responsible for source extraction, transformation, and load orchestration) designed to be built once and reused across all enterprise application workstreams.

Engineer Data Quality (DQ) Agents: Create intelligent agents for continuous monitoring of data completeness, accuracy, and consistency. Implement anomaly flagging, DQ reporting, and automated remediation triggers.

Develop the Ingestion Supervisory Agent: Build the “OneClick Ingestion” supervisor capable of generating executable ingestion code, deploying pipeline components, and managing source connectors dynamically.

Deliver Engineering Acceleration (DBT): Programmatically generate Snowflake DBT packages and build the workflow interfaces that allow engineers to review, modify, and approve the generated code.

Implement Conversational Analytics: Contribute to the “Accelerating Data Solutions” workstream by developing natural-language Q&A interfaces, automated report generation, and dynamic semantic model creation powered by Snowflake Intelligence.

Rigorous Testing & Safety Guardrails: Implement automated testing, comprehensive agent evaluation frameworks, safety guardrails, and seamless integration with human-in-the-loop approval gates.

Catalog & Governance Integration: Register every developed agent in the centralized platform catalog, maintaining meticulous documentation on capabilities, versions, inputs, and outputs.

Documentation & Knowledge Transfer: Document agent architecture designs and actively collaborate with TxDOT data engineers to ensure a smooth handoff and long-term sustainability.

Hands-on AI Engineering: Strong experience building agentic workflows, multi-agent frameworks, and utilizing LLM APIs/orchestration layers.

Snowflake & Data Ecosystem: Deep proficiency in Snowflake (Snowpark, Cortex AI, Streamlit), SQL, and Python.

Modern Analytics Engineering: Strong hands-on experience with DBT (Data Build Tool) package development, orchestration, and CI/CD pipelines.

Testing & MLOps: Experience establishing automated evaluation suites for LLMs, regression testing for data pipelines, and implementing human-in-the-loop guardrails.

Collaboration: A passion for documenting clean code, building catalogs, and conducting knowledge transfers to internal engineering teams.

Minimum Yrs of Experience, Skills, and Qualifications

7+ years in software or data engineering, including 2+ years hands-on building LLM-powered applications and at least one agent system that has run in production on live workloads. Notebook prototypes, hackathons, framework tutorials, and chat-UI wrappers do not meet this bar.

Implemented - not just configured - tool-calling agents in production: tool and function schema design, structured outputs, the orchestration loop itself, retries, timeouts, and failure handling for long-running tasks.

Built at least one production agent that generates executable code or artifacts (SQL, Python, DBT models), including automated validation and testing of generated output before it runs.

Wrote and maintained agent evaluation suites

Implemented guardrails as a builder and integration with human-in-the-loop approval gates.

Production ELT/ETL development with Snowflake and DBT: has designed, built, and operated pipelines carrying real business data. This role builds ingestion and data quality agents - data engineering depth is not optional.

Built automated data quality checks in production: completeness, accuracy, and consistency monitoring; anomaly detection; and alerting or automated remediation.

Strong Python and SQL - able to pass a hands-on coding screen without a framework doing the work.

Implementer-level experience with MCP or OpenAI-compatible function calling.

Expert-level Snowflake engineering in production environments: Snowpark, Cortex / Snowflake Intelligence, streams and tasks, security configuration, and ODS-to-business-layer modeling.

Has built or deployed MCP servers/clients or OpenAI-compatible tool interfaces in a production system - not just consumed a vendor API.

Strong Python and SQL; DBT at scale; CI/CD for data platforms; SSO (SAML/OIDC) and RBAC design.