Senior AI Engineer New Build Track Lead
Amdocs / Verizon
Remote
Note :
Responsibilities.
Lead implementation of the six agents (Orchestrator, Site Design Supervisor, Atoll RF Design via headless ATOLL through RFDA, Atoll Build Validation, FS Plot, RFDS New Build pass-through).
Lead the port-and-harden of the existing demo from POC Analysis/Demo/VerizonRFDSDemo onto Vertex AI client wiring.
Drive automated learning cycle v1 (auto-classify / auto-validate / sampled human review / governed writeback) into production.
Co-own with PID-04 the shared MCP server contracts so NB and MODS do not diverge.
Hands-on across cutover (Aug 31 Sep 5) and September soft hypercare.
Architect SME for any region-specific tuning conversation.
Must-have skills.
7+ yrs production software engineering, of which 3 yrs on LLM / agentic AI in production (not POC).
Hands-on Google Vertex AI (Agent Builder, Gemini models, RAG, embeddings, evaluation harness) built and shipped at least one production-grade agent on Vertex.
MCP (Model Context Protocol) server design and implementation built at least one production MCP server with guardrails, schemas, and rate limiting; or transferable equivalent (LangGraph tool-server, custom tool-calling protocol with hard-typed contracts).
Multi-agent orchestration patterns (state machine, retry/timeout policy, run-context persistence, compensating operations).
Python production-grade, with structured testing (pytest, type-checked, CI-integrated).
Should-have skills.
Integrated with a headless / CLI simulation tool driven by another team's API surface.
Built an adversarial validator or LLM-as-judge scoring loop with non-converge case handling.
Multi-region retrieval / KB with regional overlays.
Telecom or RF domain familiarity (will accelerate Verizon conversations).
Nice-to-have.
ATOLL / FUZE direct exposure.
Open-source contributions on agent frameworks (LangGraph, Vertex Agent Builder samples, MCP servers).
Prior demo work on the Verizon RFDS demo at POC Analysis/Demo/Verizon RFDSDemo (internal candidate).
Project overview: Verizon has launched a strategic AI transformation initiative to automate portions of its RF tower design (RFDS) process by embedding AI agents within its newly developed RFDA platform. The solution will leverage agentic workflows to analyze engineering requests, execute RF simulations, generate optimized tower design recommendations, and support human-in-the-loop decision making, with a phased adoption model progressing from AI shadowing RF engineers to fully autonomous execution.
Mission. Build, ship, and harden the New Build agent path end-to-end on Verizon's Vertex AI platform. From P2 in July through Sep 5 multi-regional cutover, this is the single named technical owner for the NB track. Becomes hypercare lead in September and product-enhancement lead Oct Nov.
Never miss an opportunity! Create an alert based on the job you applied for.