Senior AI Engineer

Amalgamated Bank of NY•New York, NY

2d•$170,000 - $200,000•Hybrid

About The Position

Amalgamated Bank is seeking an experienced Senior AI Engineer to serve as a hands-on technical lead and primary builder within its Data Governance Team. The role is responsible for designing, building, deploying, and operating governed, auditable agentic solutions on the Bank's Databricks foundation. It combines production engineering with architecture judgment and translates business needs into scalable AI capabilities that deliver measurable value in a regulated environment. The engineer will build production agents and LLM applications, shape the Bank's AI architecture, and establish secure, reusable patterns for agent tools, identity, authorization, governance, evaluation, observability, and operations. The role works closely with business stakeholders, data and BI teams, Information Security, Model Risk, Compliance, Legal, and Internal Audit. It also includes mentoring junior engineers, evaluating vendors and implementation partners, and performing other duties as directed.

Requirements

Bachelor's degree in Computer Science, Engineering, Data Science, or a related field, or equivalent experience.
7+ years of experience in software, data, or machine learning engineering, including 2+ years delivering LLM or agent-based solutions to production.
Demonstrated experience delivering agentic capabilities in a regulated industry that produced measurable business value and scaled beyond proof of concept.
Strong Python and SQL skills with hands-on production experience using Databricks, Spark, Delta Lake, Unity Catalog, and Databricks AI Gateway or a comparable LLM gateway.
Experience building and integrating MCP clients and servers, developing custom tools, implementing tool-calling workflows, and securing agent access to enterprise systems and data.
Strong prompt and context engineering skills, including model selection, evaluation design, token-usage optimization, and cost/performance trade-off management.
Knowledge of agent identity and authorization patterns, including on-behalf-of and delegated identity models.
Experience implementing guardrails for cost, accuracy, safety, observability, and end-to-end traceability of LLM and agent workloads.
Experience working directly with non-technical business teams to define scope, manage prioritization trade-offs, drive adoption, and connect AI delivery to measurable business outcomes.
Understanding of AI and data governance, model risk management, privacy, and data security in a regulated environment.
Excellent problem-solving and communication skills, with the ability to influence technical, business, risk, and executive stakeholders.

Nice To Haves

Master's degree preferred.
Financial services experience preferred.
Production experience with LangGraph, LlamaIndex, or an equivalent agentic orchestration framework.
Experience with Salesforce Data Cloud and Agentforce, MuleSoft, or Microsoft Copilot Studio.
Experience with MLflow tracing, inference tables, or comparable LLM observability tooling.
Familiarity with Google SecOps, Splunk, or comparable SIEM platforms and AI telemetry integration.
Familiarity with SR 11-7 model risk guidance and emerging AI regulatory expectations.
Databricks or Azure AI certifications.

Responsibilities

Act as a hands-on lead engineer and primary builder, writing production code and delivering agents and LLM applications while shaping the Bank's AI architecture on Databricks.
Design, build, test, deploy, and operate production agents using retrieval-augmented generation, tool calling, multi-agent, supervisory, and human-in-the-loop patterns.
Build and integrate Model Context Protocol (MCP) clients and servers; author secure custom tools and APIs; and establish reusable patterns for tool discovery, permissions, execution, and error handling.
Lead prompt and context engineering, model selection, evaluation, and token-usage optimization to improve quality, latency, reliability, and cost efficiency.
Configure and operate Databricks AI Gateway as the control point for model and agent access, including routing, access policies, rate limits, usage logging, cost controls, and guardrails.
Establish agent identity and authorization patterns, including on-behalf-of and delegated identity approaches, so agent actions are properly attributed, authorized, and auditable.
Partner with business stakeholders across credit, lending, relationship banking, and other functions to understand context and requirements and translate them into agent scope, delivery plans, and acceptance criteria.
Manage backlog priorities and trade-off decisions across business value, risk, cost, complexity, and delivery timing; drive adoption and measure and communicate business KPI impact.
Collaborate with BI engineers to architect dashboards that monitor agent cost, accuracy, usage, performance, traceability, and business outcomes.
Partner with Information Security on AI observability by identifying relevant control-plane and agent telemetry, defining signals for enterprise SIEM platforms such as Google SecOps or Splunk, and recognizing anomalous patterns that warrant investigation.
Partner with data stewards and data engineers to certify data sources and enforce classification, lineage, and entitlements through Unity Catalog.
Integrate agents across platforms using Salesforce Data Cloud and Agentforce, Microsoft Copilot Studio, and MuleSoft.
Collaborate with Information Security, Model Risk, Compliance, Legal, and Internal Audit to develop governance standards, documentation, control evidence, and testing artifacts.
Build CI/CD, automated testing, evaluation, and operational monitoring for AI workloads; mentor junior engineers; evaluate vendors and implementation partners; and perform other duties as directed.