Software Development Manager, Project Dawn
Posted Updated
As an SDM in the OAS Offers Tech org, you will own the engineering infrastructure that determines the breadth of off-Amazon merchant selection available to Amazon customers — spanning web-scale data ingestion, structured extraction, and catalog contribution.
This role sits at a genuinely hard frontier: applying large language models and hybrid extraction strategies to the inconsistent structure of millions of e-commerce websites, while maintaining the reliability and throughput requirements of a production catalog system. The right extraction approach varies by merchant, by platform, and by the maturity of the science models your team hosts and serves — and navigating those tradeoffs in real time, at scale, is one side of the challenge. The other is guiding the team through the internal complexity of Amazon's catalog — the data contracts, ingestion standards, quality thresholds, and partner system requirements that govern whether extracted merchant data successfully becomes discoverable, purchasable selection for customers. This role sits at the boundary between web-scale extraction and Amazon's catalog infrastructure, and engineering judgment on both sides is required.
You will collaborate with Science and Product partners in the Selection & Coverage pod, own the infrastructure that hosts and deploys extraction models in production, and be the engineering owner of OAS's selection coverage strategy.
Key job responsibilities
* Lead, hire, and develop a team of software development engineers across multiple workstreams, maintaining a high engineering bar and a culture of ownership
* Collaborate with Science and Product partners in the OAS Offers org, align on workstream prioritization and surface engineering tradeoffs that need leadership input
* Translate Science team model improvements into production-ready systems and establish repeatable deployment pipelines that reduce time-to-production for new model checkpoints
* Set and own the engineering delivery roadmap against committed KPI ramp targets; identify risks early and provide clear path-to-green or escalation when milestones are at risk
* Make and document active engineering tradeoffs — when to run strategies in parallel vs. sequence them, how to balance near-term funnel loss recovery against strategic infrastructure investments
* Define and maintain operational excellence standards for extraction infrastructure: availability, latency, data quality, and cost efficiency
This role sits at a genuinely hard frontier: applying large language models and hybrid extraction strategies to the inconsistent structure of millions of e-commerce websites, while maintaining the reliability and throughput requirements of a production catalog system. The right extraction approach varies by merchant, by platform, and by the maturity of the science models your team hosts and serves — and navigating those tradeoffs in real time, at scale, is one side of the challenge. The other is guiding the team through the internal complexity of Amazon's catalog — the data contracts, ingestion standards, quality thresholds, and partner system requirements that govern whether extracted merchant data successfully becomes discoverable, purchasable selection for customers. This role sits at the boundary between web-scale extraction and Amazon's catalog infrastructure, and engineering judgment on both sides is required.
You will collaborate with Science and Product partners in the Selection & Coverage pod, own the infrastructure that hosts and deploys extraction models in production, and be the engineering owner of OAS's selection coverage strategy.
Key job responsibilities
* Lead, hire, and develop a team of software development engineers across multiple workstreams, maintaining a high engineering bar and a culture of ownership
* Collaborate with Science and Product partners in the OAS Offers org, align on workstream prioritization and surface engineering tradeoffs that need leadership input
* Translate Science team model improvements into production-ready systems and establish repeatable deployment pipelines that reduce time-to-production for new model checkpoints
* Set and own the engineering delivery roadmap against committed KPI ramp targets; identify risks early and provide clear path-to-green or escalation when milestones are at risk
* Make and document active engineering tradeoffs — when to run strategies in parallel vs. sequence them, how to balance near-term funnel loss recovery against strategic infrastructure investments
* Define and maintain operational excellence standards for extraction infrastructure: availability, latency, data quality, and cost efficiency
){kind=logo}
