Software Development Engineer 3- AI & Agentic Systems

WEX•Washington, DC

2d•$121,500 - $145,500

About The Position

At WEX, a Software Development Engineer (SDE) Level 3 is a senior technical leader who designs complete services, solves complex integration problems, and drives the lifecycle of major initiatives from end to end. In this role, you will directly participate in WEX’s fast-paced AI-driven opportunities, acting as a key builder and strategist for next-generation technology. You will champion internal AI adoption by helping build AI-powered software (using tools such as Google Gemini, AWS Agent Core, Claude, and other third-party agentic integrations) while sharing education and coaching within the organization.

Requirements

Designs complete services
Solves complex integration problems
Drives the lifecycle of major initiatives from end to end
Designs and deploys multi-agent workflows within AWS Agent Core
Handles state management, memory persistence, and complex tool interactions
Drives internal education and coaching initiatives for tools such as Google Gemini Enterprise
Collaborates with external AI partners on new agentic opportunities
Designs, builds, and owns the lifecycle of entire services and cross-platform integrations
Solves complex system gaps
Drives overarching audit strategies
Makes critical trade-off decisions on data orchestration (e.g., MCP/API, Agents/Skills) based on robust data and requirements
Builds systems that handle failure gracefully by implementing retries, circuit breakers, and fallbacks
Ensures all designs are testable and maintainable long-term
Reviews code for security flaws
Ensures all services adhere strictly to WEX’s authentication standards (e.g., OAuth scopes)
Designs reliable cloud infrastructure (e.g., auto-scaling groups, failover strategies)
Proactively avoids resource over-provisioning
Owns the automated deployment package and metadata framework
Writes scripts or automation manifests to deploy configurations cleanly across sandbox tiers
Prevents environment drift
Builds and optimizes CI/CD pipelines
Focuses heavily on reducing build times and tightening feedback loops
Solves complex bugs (intermittent issues, race conditions, memory leaks)
Ensures the root cause is fixed permanently
Owns Incidents for your services during outages
Creates step-by-step triage runbooks and playbooks that empower Tier 1/2 helpdesk teams or junior analysts to resolve common platform errors without escalation
Partners with Product Managers who define Functional requirements by owning the Non-Functional requirements (Performance, Security, Scalability) needed to meet business KPIs
Works with PMs to prioritize the backlog
Explicitly negotiates the inclusion of "Tech Debt" and "Refactoring" by explaining long-term business risks
Negotiates "Must Haves" vs. "Nice to Haves" to ensure sprint achievability
Offers alternative solutions that respect time and budget constraints
Enforces strict quality standards during code reviews with empathy
Rejects sloppy or brittle "quick fixes"
Sets the organizational standard for clean design patterns
Proactively mentors SDE 1s and 2s
Delegates work specifically to help them grow
Facilitates data-driven discussions to resolve engineering disagreements
Effectively de-escalates "religious wars" about technology
Ensures "quiet" voices are heard in design meetings
Rapidly learns and applies new languages or frameworks
Leads knowledge-sharing sessions
Seeks out critical feedback
Context-switches effectively across workstreams when project parameters change

Nice To Haves

Using tools such as Google Gemini, AWS Agent Core, Claude, and other third-party agentic integrations

Responsibilities

Design and deploy multi-agent workflows within AWS Agent Core, effectively handling state management, memory persistence, and complex tool interactions.
Help drive internal education and coaching initiatives for tools with as Google Gemini Enterprise; collaborate with external AI partners on new agentic opportunities.
Design, build, and own the lifecycle of entire services and cross-platform integrations. Solve complex system gaps and drive overarching audit strategies.
Make critical trade-off decisions on data orchestration (e.g., MCP/API, Agents/Skills) based on robust data and requirements.
Build systems that handle failure gracefully by implementing retries, circuit breakers, and fallbacks. Ensure all designs are testable and maintainable long-term.
Review code for security flaws and ensure all services adhere strictly to WEX’s authentication standards (e.g., OAuth scopes).
Design reliable cloud infrastructure (e.g., auto-scaling groups, failover strategies) while proactively avoiding resource over-provisioning.
Own the automated deployment package and metadata framework. Write scripts or automation manifests to deploy configurations cleanly across sandbox tiers, preventing environment drift.
Build and optimize CI/CD pipelines, focusing heavily on reducing build times and tightening feedback loops.
Capable of solving complex bugs (intermittent issues, race conditions, memory leaks), ensuring the root cause is fixed permanently.
Own Incidents for your services during outages. Create step-by-step triage runbooks and playbooks that empower Tier 1/2 helpdesk teams or junior analysts to resolve common platform errors without escalation.
Partner with Product Managers who define Functional requirements by owning the Non-Functional requirements (Performance, Security, Scalability) needed to meet business KPIs.
Work with PMs to prioritize the backlog, explicitly negotiating the inclusion of "Tech Debt" and "Refactoring" by explaining long-term business risks. Negotiate "Must Haves" vs. "Nice to Haves" to ensure sprint achievability.
Offer alternative solutions that respect time and budget constraints.
Enforce strict quality standards during code reviews with empathy ("Check tone"). Reject sloppy or brittle "quick fixes" while setting the organizational standard for clean design patterns. Proactively mentor SDE 1s and 2s, delegating work specifically to help them grow.
Facilitate data-driven discussions to resolve engineering disagreements, effectively de-escalating "religious wars" about technology and ensuring "quiet" voices are heard in design meetings.
Rapidly learn and apply new languages or frameworks, lead knowledge-sharing sessions, seek out critical feedback, and context-switch effectively across workstreams when project parameters change.