Software Development Manager, Alexa Sensitive Content Intelligence (ASCI)

About the role

The Alexa Sensitive Content Intelligence (ASCI) team builds the Responsible AI platform that keeps Alexa+ content safe worldwide. As a Software Development Manager you will lead an engineering team that delivers production‑grade ML inference infrastructure and content‑safety services for both cloud‑based and on‑device moderation.

Key responsibilities

• Lead and grow a high‑performing engineering team delivering ML inference and content safety services for Alexa+ Responsible AI.
• Own inference optimization, including GPU/CPU serving strategies for cloud and model compression/distillation for on‑device deployment.
• Design and operate a one‑click continuous model deployment pipeline supporting encoder, decoder, and encoder‑decoder models with zero‑downtime rollouts, automated validation, and rollback.
• Architect scalable real‑time and batch serving systems handling multi‑modal content (text, image, video, LLM invocations) for worldwide moderation.
• Drive on‑device moderation capabilities, optimizing models for latency, memory, and power constraints while maintaining accuracy.
• Ensure operational excellence with high availability (99.95%+), low latency, cost efficiency, and multi‑region reliability.
• Partner with applied scientists on model architecture trade‑offs and with platform teams on traffic migrations, capacity planning, and device integration.
• Mentor and develop SDEs/MLEs with a focus on emerging ML hosting and infrastructure optimisation technologies.

Required profile

• Proven experience leading engineering teams that build and operate ML inference services.
• Strong collaboration skills with applied scientists, product managers, and platform engineering groups.
• Ability to define technical strategy, make architecture decisions, and communicate status and risks to senior leadership.

Required skills

• ML inference infrastructure
• GPU/CPU serving optimization
• Model compression and distillation
• Continuous deployment pipelines
• Zero‑downtime rollouts
• Automated validation and rollback mechanisms
• Real‑time and batch serving systems
• Multi‑modal content processing
• On‑device deployment
• Cloud and edge deployment
• Capacity planning and traffic migration