Senior SW AI Engineer

Visa

Confirmed live today High trust

Hybrid

Quick summary

Work type: Hybrid
Location: Austin, TX
Salary: $110,700–$171,800 / yr
Employment: Full-time
Posted: 25 days ago
Freshness: Confirmed live today
Nearby: 99+ roles within 25 mi

Market check

Salary context

Below market

How this pay compares to similar roles

Similar $190k

This role $141k

$96k $245k

below market most similar roles pay here above market

This role pays less than 84% of similar roles. Most pay $149,500–$230,887 — the blue band above. At the midpoint, this role pays about $141k versus about $190k for comparable roles.

Based on 240 similar postings.

Employer

About Visa

Visa Inc. is the world''s largest payment network, enabling digital payments for consumers, merchants, financial institutions, and governments through its global processing network and branded card products. Industry: Payments Technology & Financial Services

Visa currently has 114 open roles on FindRole.

Listed pay typically runs $161,950–$259,450 across 110 roles with salary data.

Most-posted roles

View all roles at Visa

At a glance

TL;DR · Senior SW AI Engineer

Role Posting Log in to save

JOB TITLE: Sr. SW AI Engineer The Sr. SW AI Engineer joins an AI-first engineering team to build production-grade agentic systems powered by large language models on on-premise infrastructure. This role involves designing and implementing multi-agent orchestration frameworks, retrieval-augmented generation pipelines, and tool-calling integrations. You will develop and maintain Spring Boot microservices using Java 21, LangGraph, LangChain4j, and PostgreSQL with pgvector. Key responsibilities include building ReAct agent loops, implementing Model Context Protocol tool servers, and managing LLM inference gateways via REST APIs. You will also author Helm charts and Jenkins CI/CD pipelines for Kubernetes deployments. The work focuses on solving fintech sector challenges by creating scalable, secure AI solutions that include agent memory architectures, confidence scoring, and structured audit trails for traceability and explainability.

Skills

Java Spring Boot LangChain LangGraph LLM RAG PostgreSQL pgvector Kubernetes OpenShift Helm Jenkins CI/CD REST API Prompt Engineering Microservices AOP

What you'll do

Develop and maintain production-grade multi-agent orchestration frameworks using LangGraph or LangChain.
Build and optimize RAG pipelines including document chunking, embedding generation, and vector storage in PostgreSQL.
Implement ReAct agent loops and supervisor patterns to manage complex LLM workflows at enterprise scale.
Develop Spring Boot microservices to host AI engines and integrate with internal data sources via REST APIs.
Create and maintain Model Context Protocol (MCP) tool servers to expose enterprise data for AI agents.
Build automated CI/CD pipelines and Helm charts to deploy containerized AI services on Kubernetes or OpenShift.
Design audit trails and observability tools to track LLM reasoning, tool calls, and performance metrics.
Create comprehensive test suites for non-deterministic AI components to identify prompt drift and ensure output quality.

What we're looking for

2 or more years of work experience with a Bachelor’s Degree or an Advanced Degree (e.g., Masters, MBA, JD, MD, or PhD).
Experience in developing backend services in Java (Java 17+ preferred; Java 21 a strong plus).
Familiarity with REST API design, Spring Boot, and JPA/ORM-based data access patterns.
Foundational understanding of LLM APIs including prompt construction, token limits, and response parsing.
Experience building RAG pipelines involving document chunking, embedding models, and vector databases like PostgreSQL with pgvector.
Hands-on experience building agentic AI systems using LangGraph, LangChain, LangGraph4j, or LangChain4j (preferred).
Experience implementing the ReAct prompting pattern in a production or near-production LLM application (preferred).
Experience deploying containerized workloads on Kubernetes or OpenShift using Helm charts and Jenkins CI/CD pipelines (preferred).

Save