Software Architect - Agentic AI +Telecom

Experience: 10 years

Domain: Agentic AI Telecom

Location: Dallas Texas

What You'll Be Doing:

• Creating framework applications/features and developing POCs for new feature requests in Verizon's agentic ecosystem.
• Owning and driving high-level design, development, business stakeholder engagement, Platform Architecture discussions, and external application team SMEs.
• Designing and building agentic platform capabilities across frontend, backend microservices, and AI agent services — including conversational workflows, intelligent intake, content generation, and campaign orchestration.
• Leading end-to-end solution design spanning React-based web applications, Java Spring Boot services, Python FastAPI/LangGraph agents, and cloud-native autonomous services.
• Building and maintaining RAG pipelines across Google Cloud Platform, Oracle RDS, and PostgreSQL — and owning validation of retrieval quality and agent output accuracy.
• Monitoring and comparing agent performance across non-prod and prod environments — latency, accuracy, retrieval quality, and LLM output consistency — to validate readiness before production rollout.
• Delivering incremental, production-ready builds with feature flags, safe rollout patterns, and cross-repo coordination.
• Setting standards for architecture, scalability, design, and runtime performance — including async agent flows, parallel API integration, RAG optimization, and frontend/backend performance tuning.
• Partnering with SRE, platform, AID, and distributed engineering teams (Verizon India and US) to unblock delivery, production changes, and integration with enterprise systems.

What We're Looking For:  

• Bachelor's degree with 10 or more years of work experience.
• Experience with large scale projects.
• 3 or more years of experience as an Architect or Senior Developer in a large and complex enterprise setting.
• Two or more years of specific work experience with building Agentic capabilities.
• Strong hands-on experience with Java and Web development technologies — Spring Boot, REST APIs, React, and modern frontend patterns.
• Experience working with Microservices architecture and event-driven systems (Kafka, BFF pattern).
• Experience with Python agent/AI frameworks (LangChain, FastAPI) and LLM performance optimization.
• Hands-on experience with LangGraph (or equivalent agent orchestration frameworks) for building conversational and multi-step agent workflows.
• Hands-on experience with RAG development — building retrieval pipelines on Google Cloud Platform, Oracle RDS, and PostgreSQL, embedding strategies, hybrid search, and LLM-grounded knowledge systems.
• Ability to own validation of RAG and agent results — accuracy checks, retrieval quality assessment, and output verification before production rollout.
• Experience monitoring and comparing agent performance between non-prod and prod — tracking latency, token usage, retrieval hit rates, response quality, and regression across environments.
• Experience with Docker containerization and Helm-based Kubernetes deployments.
• Strong understanding of end-to-end enterprise application flows — API design, workflow orchestration, data integration, and production deployment patterns.
• Ability to work independently on implementations while participating actively in architectural and design discussions.
• Certifications (Optional — e.g. Google Cloud Platform/AWS Architect, Kubernetes CKA).

Even Better If You Have:

• Experience in the telecom domain and understanding of marketing campaign strategies — offer design, audience targeting, promotional campaigns, and launch planning.
• Experience with Adobe Campaign and AEM (Adobe Experience Manager).
• Experience with Cloud platforms (Google Cloud Platform, AWS, Cloud Run or equivalent serverless).
• Experience integrating with large enterprise systems — catalog/pricing APIs, workflow tools (Jira, Workfront), and content platforms.
• Experience coordinating production changes with SRE and cross-region distributed teams.
• Excellent communication skills.