Responsibilities
Build the solution: Own end-to-end technical delivery of agentic systems, from source-system integration through agent design, development, evaluation, and production deployment.
Integrate AI systems with our source systems (Salesforce, Databricks, Splunk, internal APIs, business applications). Handle agent harness and orchestration.
Handle the operational handover to the business function and any necessary support transition.
Collaborate across teams: Partner closely with data engineering, platform, security, and business teams to align on requirements, dependencies, and integration points.
Engage stakeholders: Gather requirements directly from business functions, communicate technical trade-offs clearly, and keep stakeholders informed on progress, risks, and timelines.
Uphold quality and reliability: Establish and maintain best practices for code quality, testing, evaluation, monitoring, and observability of deployed AI systems.
Contribute to the team's technical growth through knowledge-sharing and help shape engineering standards and reusable patterns.
Stay current: Continuously evaluate emerging Gen AI, LLM, and agentic frameworks, and recommend tools and approaches that improve delivery speed and solution quality.
Qualifications
5+ years of ML / AI engineering, with at least 2 years building production AI / Agentic systems.
Hands-on with at least one Agentic framework (LangGraph, CrewAI, or custom) and an LLM provider's production tooling APIs.
Fluency in Python.
Track record of shipping fast: has examples of taking an AI system from idea to production in weeks, not quarters.
Comfortable working directly with business stakeholders without a product manager intermediary on every interaction.
Nice to have
Payments domain knowledge or experience in another regulated industry.
Production experience with Snowflake, Splunk, or similar enterprise data and observability platforms.
Open-source AI tooling contributions or technical writing.
Has built evaluation harnesses for LLM systems beyond simple accuracy metrics (e.g. hallucination scoring, agent trajectory eval).
