Temporary Watchdog Jobs (NOW HIRING)

We’re building an autonomous site-level connectivity intelligence platform for enterprise environments. This role owns the on-device agent and key parts of the device-facing middleware, and also contributes broadly to our Python backend as we scale from pilots to fleets.

You’re a “full-stack systems” engineer: comfortable shipping reliable software on edge devices and building clean backend services, APIs, and data pipelines in Python.

• Edge agent runtime on OpenWrt/Linux:
  • service lifecycle, watchdog, structured logging, resource controls
• Telemetry & probes:
  • device health (CPU/mem/disk/uptime/temp where available)
  • network health (interfaces, IP, routes, gateway reachability, DNS config)
  • internet probes (ping/DNS/HTTPS timing, loss/latency/jitter derived)
  • private cellular telemetry: RSRP/RSRQ/SINR, RAT, band/channel, cell ID/PCI, attach/detach, SIM state
  • Wi‑Fi telemetry: passive scans (SSID/BSSID, RSSI, channel, band, density), link stats when associated (best‑effort)
• Store‑and‑forward reliability:
  • local spooling (SQLite or equivalent), bounded storage, backoff/retry, idempotent batching, backfill after outages
• snapshot capture on triggers, consistent schemas, versioning, compatibility discipline
• Hardware abstraction layer to support multiple device SKUs over time

• Device registry & fleet health:
• identity, capability discovery, last‑seen/heartbeat, data completeness scoring
• Backend product primitives:
  • event processing, correlation inputs, incident objects, site‑level rollups
  • data models and query patterns supporting “evidence‑first” RCA
• Ops tooling:
  • dashboards, ingest failure triage, replay tools, offline detection

• Architect and implement the edge agent to be production‑grade (unattended, self‑healing, resilient to flaky power/backhaul).
• Own devicecloud schemas and compatibility strategy.
• Build and operate core Python backend services and data pipelines.
• Establish testing and release discipline for edge + backend:
  • unit/integration tests, network fault injection, replayable fixtures
• Drive security/reliability best practices:
  • per‑device identity, TLS everywhere, least privilege, safe upgrade path

• Strong experience building production Python systems (services, APIs, workers).
• Experience with telemetry pipelines: batching, retries, idempotency, persistence, backpressure.

• Cellular: ModemManager, QMI/MBIM, AT commands, SIM/attach state.
• Wi‑Fi: iw, nl80211 concepts, scan/link stats.
• Backend infra: Postgres, Redis, queues/workers, time‑series/event storage.
• Device fleet security: keys/certs, enrollment, signed updates.

• Ownership is real: if it breaks in the field, we fix it—fast—and we harden it so it doesn’t recur.
• Bias to shipping: meaningful progress weekly; small PRs; tight feedback loops.
• High standards on reliability: correctness > cleverness; explicit contracts; versioned schemas.
• No “throw it over the wall”: edge + backend are one system; we optimize the whole loop.
• Low ego, high velocity: debate the idea, not the person; be direct and respectful.

• 30+ day unattended stability on devices in real environments.
• No data loss during outages (bounded spool + backfill works).
• Ingestion is resilient and idempotent; device registry and fleet health are reliable.
• Evidence bundles and symptom events are consistent and useful to downstream RCA.