About us

• Attribution Engineer — Gravity
• Location: San Francisco, CA (on-site) Company: Gravity (trygravity.ai)
• About Gravity
• Gravity is an AI-native ad network that serves millions of ads per day to AI platforms. We match ads to conversational context in real time using ML scoring models and LLM-powered ad generation. We raised our seed round at a $75M valuation and are growing quickly.

Role

• Attribution is our #1 engineering priority. Accurate attribution data is what allows our system to learn and improve — without it, we can't optimize our models or prove ROI to advertisers. It's foundational to everything else we build.
• We need someone who can own the full attribution system end-to-end. You'll work directly with the CTO and a small, senior engineering team. The problems are hard: real-time event ingestion, cross-device identity resolution, deterministic and probabilistic attribution methods, and building the data layer that our ML pipeline depends on.
• What You'll Work On
• You'll own Gravity's attribution pipeline end-to-end. That includes pixel infrastructure on advertiser sites, server-side integrations (such as conversion APIs), event ingestion and deduplication, identity resolution, delayed attribution for conversions that happen after ad exposure, and conversion reporting.
• You'll also work on connecting attribution signals back into our user profiling and ad targeting systems in real time. The attribution data you produce feeds directly into our ad engine, which is evolving toward a DLRM-inspired neural ranker with multi-task prediction heads. Accurate, timely data from attribution is a prerequisite for that system to work.
• What We're Looking For
• Someone who has built or substantially improved an attribution system at an ad network, ad platform, or adtech company. You understand multi-touch attribution, view-through vs. click-through modeling, and the tradeoffs between deterministic and probabilistic approaches. You've dealt with event data at scale and you know what it takes to make conversion numbers trustworthy.
• Strong data engineering fundamentals — event-driven architectures, real-time pipelines, queue-based systems. You think carefully about idempotency, deduplication, and failure handling.
• Experience with or interest in ML-adjacent work is a plus. The attribution system feeds directly into our ranking models, so understanding how downstream systems consume your data makes you more effective.
• Why Gravity
• Small team, real scale, hard problems. Attribution for AI-native platforms hasn't been solved before — you'd be building it, not maintaining someone else's system. Full ownership of a critical system, working directly with the founders. Equity is part of the package.
• Tech Stack
• Python, Kafka, Machine Learning, AWS