This role is four days onsite, with the flexibility to work from home one day per week M&T Work Model - 4 Days In-Office | 1 Day Work From Home
The Bank sponsors individuals for TN and H-1B transfers on a case by case basis. Please note that this position is not open to anyone on an F-1 student visa including those eligible for CPT/OPT or the Stem OPT extension.
Overview:
Hands-on technical lead within an application team, delivering production-grade agentic workflows, including mainframe modernization, within the bank's established standards and controls, and moving the team from AI-assisted coding to spec-driven, agent-led delivery.
Primary Responsibilities :
Application Delivery & Technical Leadership
•Identify and prototype agentic use cases across the team's SDLC: requirements, coding, testing, review, modernization, documentation, and incident support.
•Align delivery with business priorities and Enterprise Architecture standards.
Agentic Engineering Excellence
•Build agentic workflows using Model Context Protocol (MCP) tools, agent-to-agent (A2A) communication, multi-agent orchestration, context and memory management, and human-in-the-loop checkpoints.
•Optimize quality, reliability, latency, and cost, including model selection by task fit and data classification.
•Release prompts, agent configurations, and models through bank CI/CD with automated regression evaluation.
•Coordinate parallel background coding agents whose pull requests engineers review and approve.
Specification-Driven Development (SDD)
Apply the bank's SDD process, with versioned specifications (specs) as the source of truth for agent-generated work:
•Specify – Capture intent, business rules, acceptance criteria, and non-functional and control requirements with product, engineering, and risk partners.
•Plan – Translate the spec into a design (architecture, interfaces, data contracts) aligned to enterprise standards.
•Decompose – Break the design into small, testable, traceable tasks.
•Implement – Direct agents to generate code, tests, and documentation, with human review.
•Validate – Verify against the spec through tests and evaluations; change the spec first, then regenerate.
•Reverse-engineer COBOL/JCL into SME-validated as-is specs that drive to-be design and functional-equivalence testing.
•Maintain application specs within bank templates, guardrail ( constitution ) files, and context engineering standards, with audit-ready traceability from spec to code, tests, and change records.
Evaluation, Measurement & Quality Engineering
•Apply bank evaluation frameworks with application-specific baselines, golden datasets, and red-team tests covering quality, reliability, cost, and developer impact.
•Monitor production with end-to-end tracing and online evaluation to catch drift, silent tool failures, and cost overruns.
•Ensure deployment readiness and substantive review of AI-generated code per bank standards.
•Report outcomes using DORA software delivery metrics, cycle time, test coverage, rework rate, and developer satisfaction.
Risk, Security & Governance
•Implement bank guardrails: least-privilege permissions, secrets handling, prompt injection defenses, data classification, audit logging, approval gates, sandboxing, blast-radius limits, kill switches, and rollback.
•Register agents as non-human identities with owners and access recertification.
•Use only approved models, MCP servers, and plugins; route new ones through AI intake review and keep AI bill of materials (AI-BOM) entries current.
•Comply with bank risk, regulatory, change, security, data protection, and operational standards and AI governance frameworks (SR 11-7, NIST AI RMF, ISO/IEC 42001); escalate risks and close audit items on time.
Collaboration & Continuous Improvement
•Adopt enterprise patterns, playbooks, and approved catalog assets (agent skills, tools, AGENTS.md files); contribute reusable components and lessons learned through established channels.
•Share AI tooling and standards gaps with Enterprise Architecture, platform, and governance teams.
•Communicate and document designs, decisions, risks, and tradeoffs, including AI limitations and when agents are not the right fit.
•Coach engineers on effective agent use and bank standards; participate in the enterprise community of practice.
Standard Expectations
•Support a culture of belonging, reflect the M&T Bank brand, maintain internal controls, and complete other assigned duties.
Scope of Responsibilities
•Accountable, without direct reports, for the application's agentic workflows (design, delivery quality, evaluation, guardrail adherence, operational readiness, production support, and adoption), working with product, architecture, platform, security, operations, risk, and audit partners.
Education & Experience Required:
•Associate’s degree and a minimum of 7 years’ systems analysis and/or application development work experience or Bachelor's degree and a minimum of 5 years' systems analysis and/or application development work experience. In lieu of a degree, a combined minimum of 9 year’s education and/or relevant work experience, including a minimum of 5 years’ system analysis and/or application development work experience.
•Production experience with LLM-based or agentic systems: tool calling, retrieval, context engineering, and prompt design.
•Hands-on with MCP or equivalent tool integrations, A2A, and multi-agent orchestration.
•Evaluation design for non-deterministic systems (harnesses, golden datasets, metrics).
•Writing precise, testable specifications and acceptance criteria.
•Python and/or Java/TypeScript; comfortable in large legacy codebases.
•Strong software design, testing, code review, and CI/CD (GitLab or equivalent) fundamentals.
•Experience leading technical workstreams on production systems and driving adoption through influence.
•Track record of delivering within enterprise SDLC, resiliency, security, and change management controls.
Education & Experience Preferred:
•Enterprise AI coding platforms such as GitLab Duo, and SDD tools such as GitHub Spec Kit or Kiro.
•Legacy modernization at scale (framework upgrades, language migrations, test backfill, COBOL/JCL reverse engineering).
•Regulated-industry experience, especially financial services model risk, audit, and data governance.
•Forward deployed, solutions, or consulting-style engineering experience.
•Open source or technical writing contributions.
We support our team members with generous benefits.
•Competitive compensation
•Health, welfare, and retirement benefits
•401(k) match at 5%
•Work-life balance and flexible work arrangements
•Banking Officers start with 25 days PTO plus 12 paid holidays
•40 hours paid volunteer hours per year
•Much more. For details, see: M&T Benefits Overview
#LeadForwardDeployedEngineer #ForwardDeployedEngineer #AgenticEngineering #AgenticAI #AIEngineer #GenerativeAI #LLM #MCP #MultiAgentSystems #SpecDrivenDevelopment #MainframeModernization #COBOL #LegacyModernization #Python #Java #TypeScript #CICD #GitLab #AIGovernance #FinancialServices #BankingJobs #MandTBank #NowHiring #TechJobs #SoftwareEngineering
M&T Bank is committed to fair, competitive, and market-informed pay for our employees. The pay range for this position is $116,400.00 - $194,000.00 Annual (USD). The successful candidate’s particular combination of knowledge, skills, and experience will inform their specific compensation. The range listed above corresponds to our national pay range for this role. The specific pay range applicable to you may vary based on your location.
Location
Clanton, Alabama, United States of America