Senior Software Engineer II, Agentic AI Platform

LVT•Seattle, WA

About The Position

LVT is redefining how businesses operate in the physical world, moving beyond traditional security solutions to deliver AI-driven, actionable intelligence that makes sites smarter, safer, and more secure. Since pioneering our first mobile, solar-powered units, our commitment to scrappy, hands-on innovation has made us an established leader and one of the fastest-growing companies in intelligent site technology. We are building the next generation of solutions—from our physical units in the field to a powerful Agentic AI platform—that allows our customers to gain unprecedented visibility and control over safety, compliance, and operations. This is your chance to join a cutting-edge team that isn't just watching the world change, but actively building the technology that is changing it. We’re a team that’s focused on growth and innovation, and we’re proud that our crew, products, and leadership are being recognized for it. A Top-Tier Growth Company: Named one of the Financial Times’ Fastest Growing Companies 2025 and #10 on the Inc. 5000 Rocky Mountain Regional list for 2025. Innovative Leadership: Our CEO, Ryan Porter, was named an EY Entrepreneur of the Year 2025, and our CTO, Steve Lindsey, was inducted into the Silicon Slopes CTO Hall of Fame in 2024. Product & Software Excellence: We were named one of The Software Report’s Top 100 Software Companies of 2023 and are a winner of the Security Today Govies Award for 2025. LVT is building an agentic AI platform that enables autonomous agents that perceive real-world environments, reason over them, and act. Agents operate on an event pipeline that feeds them what's happening across tens of thousands of edge devices and carries their decisions back out to the field. This role builds and hardens the cloud backbone of that platform the event pipeline agents consume from and act through, and the context stores they query. You'll own the cloud-side reliability, throughput, and scale of the event pipeline, the router/dispatcher that fans events to intelligence services, the projections that make events queryable (including the agent-facing, MCP-exposed context layer), and the observability that keeps it healthy under load. You will build the platform that lets agents run reliably at fleet scale. You'll partner closely with the edge team that produces events, the data team that owns the lakehouse you feed, infrastructure/operations who run the managed log and broker, and the engineers building the agent and CV workflows that consume your platform.

Requirements

6+ years building and operating backend or distributed systems in production at scale, with strong systems and API design experience.
Hands-on with Kafka or a comparable log/streaming system; topic and partition design, consumer-group semantics, delivery guarantees, replay, and schema-registry integration.
A track record of hardening services for production including SLOs, metrics and alerting, dead-letter and failure handling, and the operational maturity to own on-call trade-offs.
Comfortable choosing and operating different stores per access pattern (document/relational hot stores; object storage / lakehouse for cold and analytical).
Strong in Go and/or Python; experience building APIs and large-scale backend services.
AWS and Kubernetes, with a bias toward open, cloud-portable components.
Bachelor's or Master's in Computer Science, Engineering, or a related field, or equivalent practical experience.

Nice To Haves

Building agent-facing or MCP data-access layers, governed query interfaces over operational data for autonomous consumers.
Integrating agent or CV runtimes onto an event pipeline as pluggable consumers.
Stream-processing frameworks (Benthos/Bento, Flink, or similar).
Schema and contract tooling (protobuf/Avro, schema registry).
Lakehouse producer experience (Iceberg / S3, CDC); MQTT / IoT ingest at scale (EMQX or comparable).

Responsibilities

Harden the event pipeline: Make the cloud pipeline production-grade consumer reliability, delivery and ordering guarantees, replay, idempotency, dead-letter handling, and schema/contract enforcement at the boundary.
Serve the agent layer: Make the dispatch path to agent and CV workflows reliable, and expose the event projections via MCP / agent-facing query interfaces that agents read for context, so the agentic platform has a dependable substrate to run on.
Scale ingress and projections: Stand up and scale the materialized stores the pipeline serves from (operational hot store, cold capture into the lakehouse), keeping each projection fit to its access pattern and rebuildable from the log.
Observability and operability: Own the signals that make the pipeline supportable, own dashboards, alerting, and feature flags.