Senior Software Engineer - Data Platform

Aira Technologies; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Direct access to founders

Multi‑agent systems in production

Global tier‑1 operators

High autonomy

Job summary

Aira Technologies is seeking a senior data systems engineer to transform raw cellular signals into an intelligence layer used by autonomous network agents. You will build streaming pipelines, model data, and APIs powering real‑time decision making.

You will collaborate with ML researchers, wireless engineers, and product teams to ship analytics to tier‑1 operators, with high autonomy and ownership over the full data stack.

Qualifications

• 5+ years building production systems with data at the center.
• Deep SQL and understanding of query planning, partitioning, and warehouse cost/performance tradeoffs.
• Strong Python; production experience with Spark and workflow orchestration (Airflow, Dagster, or similar).
• Backend API experience (FastAPI, Flask, or equivalent) serving data to applications and models.
• Time‑series or event data at scale: schema evolution, late arrivals, idempotency, backfill strategy.

Responsibilities

• Build the data backbone transforming signals from millions of towers into the intelligence layer for real‑time agent optimization.
• Develop streaming pipelines and a gold data layer used by dashboards and models.
• Serve typed, paginated APIs over the gold layer for UI and agents, ensuring fast responses.
• Collaborate with ML researchers, wireless engineers, and product teams to ship analytics to tier‑1 operators.

Skills

SQL

Python

Backend APIs

Time-series data

Production systems

Tools

Spark

Airflow

Dagster

FastAPI/Flask

Snowflake

Job description

We build AI agents that run cellular networks.

Mobile operators spent billions on 5G but can't extract value—vendor lock-in blocks innovation, monolithic systems don't scale, operational complexity multiplies. We deploy multi-agent systems with fine-tuned LLMs that autonomously optimize network performance, generate operational tools, and execute operator intent. Applied AI at planetary scale: millions of towers, petabytes of telemetry, real-time inference affecting billions of users.

Founded by Qualcomm and Intel veterans. Team from Stanford, Berkeley, Cornell. Backed by AT&T Ventures, Intel Capital, In-Q-Tel, NeoTribe, Acrew, and more.

The Role

Build the data backbone that makes autonomous networks possible. You'll transform raw signals from millions of cell towers into the intelligence layer our AI agents use to optimize in real-time - streaming pipelines that feed agent decision‑making, models that power operator dashboards, feature stores that enable inference at scale. Your work directly determines what our agents can see and how fast they can respond.

You'll collaborate directly with ML researchers tuning models on your data, wireless engineers defining what matters, and product teams shipping analytics to tier‑1 operators. Ship fast, iterate, own the full stack from ingestion to insight.

What You’ll Build

• Config‑driven ingestion from multi‑vendor RAN element managers into cloud object storage and Snowflake - bronze through gold
• A modeled KPI layer at cell, carrier, and market granularity that agents and dashboards both query, with contracts stable enough to build on
• Serving APIs over the gold layer - typed, paginated, cached, and fast enough for interactive queries at cell granularity, consumed by both the UI and our agents
• Data quality, freshness SLAs, lineage, observability, and cost controls for datasets a tier‑1 NOC depends on

What We’re Looking For

• 5+ years building production systems with data at the center of them
• Deep SQL, plus real understanding of query planning, partitioning, and warehouse cost/performance tradeoffs
• Strong Python; production experience with Spark and workflow orchestration (Airflow, Dagster, or similar)
• Backend API experience (FastAPI, Flask, or equivalent) serving data to applications and models
• Time‑series or event data at scale: schema evolution, late arrivals, idempotency, backfill strategy
• Bonus: telecom/RAN data (3GPP NRM, PM/CM/FM), dbt, Iceberg/Delta, or building serving layers that LLM agents query

No telecom background required - we’ll teach you the domain; we can’t teach you judgment about production data.

Why Aira

• Direct access to founders, partners and researchers
• Work on multi‑agent systems and fine‑tuned LLMs in production before most companies figure out prompts
• Partner with tier‑1 mobile operators globally
• High autonomy, real ownership, visible impact

We’re building the intelligent network. Help us ship it.