Senior Software Engineer, Agentic Ads Experience

22h•$190,800 - $267,100•Remote

About The Position

Reddit's Ads business is at an inflection point, and generative AI is reshaping how advertisers plan, build, launch, measure, and optimize campaigns. The Agentic Ads Experience team is building the next generation of advertiser-facing interfaces: campaign assistants, creative and ad copy generation, one-click campaign creation, audience discovery, conversational analytics, and optimization recommendations. This is a unique opportunity to join a fast-moving team shipping generative AI/agentic systems into a high-growth Ads product, where every launch has direct advertiser and revenue impact.

Requirements

5+ years of experience as a software engineer, with a track record of shipping and operating production systems.
Hands-on experience building generative AI applications: structured outputs, function/tool calling, prompt and context design, RAG, model routing, and handling model failure modes.
Genuine enthusiasm for agents and LLMs — you follow the space closely, have opinions on the tools, and can talk shop about the evolving ecosystem.
Strong systems fundamentals: APIs, distributed systems, queues, data stores, service ownership, auth, idempotency, retries, and observability.
Experience defining evals and measurement for AI systems — offline datasets, online metrics, and knowing when and how to measure a launch quality bar.
Fluent with AI coding tools and code-assistant workflows as part of your day-to-day engineering practice.
Excellent communication skills. You can collaborate effectively in a remote environment and explain complex AI system trade-offs to technical and non-technical audiences.
Extensive software development experience in one or more general purpose languages: Python, Golang, Rust, Java, JavaScript, etc.

Nice To Haves

Experience in Ads, marketing technology, marketplaces, growth, recommendations, search, or analytics.
Experience building AI systems where mistakes are expensive: customer-facing products, regulated workflows, policy review, payments, or high-scale infrastructure.
Experience building internal workflow automation for sales, ops, support, policy, trust, or engineering productivity.

Responsibilities

Design and ship advertiser-facing AI features and platform (campaign assistants, creative generation, audience discovery, performance analytics, optimization recommendations) end-to-end, from prototype to GA.
Make the right architectural calls on if and how to use generative AI, whether it’s a single-shot LLM call, a workflow, a skill, a retrieval system, an agent loop, human review, or no AI at all.
Build for the failure modes of LLMs — structured outputs, tool/action boundaries, retries, cost controls, tracing, evals, and human-in-the-loop where it matters.
Define offline and online evals, instrument traces, run A/B tests, and connect model and product metrics to advertiser and business outcomes.
Design for data isolation, permissioning, auditability, and policy compliance when AI systems take action on behalf of advertisers.
Partner closely with PM, Design, Data Science, ML Platform, Trust, and sales/ops to turn ambiguous problems into concrete, shippable systems. Collaborate with other product teams on how best to expose their services and capabilities to the platform’s centralized agent(s).
Write clear design docs, explain trade-offs to non-engineers, and raise the bar on how the org ships gen AI features.