Backend Software Engineer Camera & Photos Tools & AI Team

At Apple new ideas have a way of becoming extraordinary products and experiences very quickly. Bring your passion and dedication to your job and theres no telling what you could Camera u0026 Photos Tools u0026 AI team is a tight-knit engineering team building the internal tools that power how the Camera Photos and Image Quality teams measure evaluate and improve the imaging experience on Apple products. Our software sits at the center of some of Apples most demanding imaging workflows: it captures and catalogs enormous volumes of images and videos orchestrates long-running analyses that characterize camera performance and surfaces the results to the engineers and scientists who tune the hardware and software behind every photo our customers move quickly care about the craft and turn ambiguous problems into reliable well-designed systems. Youll own backend services and data infrastructure end to end from Python REST APIs to the model-serving infrastructure behind our AI-native tooling partnering with engineering science and quality teams across Camera Photos and Image Quality. As AI capabilities advance rapidly our team is actively building AI-native tooling from integrating multimodal and vision models into image quality workflows to designing LLM-powered interfaces that let engineers query and interpret large datasets in natural language. We want someone who doesnt just call a hosted API but who can design deploy and operate the serving layer underneath it and who holds AI-powered features to the same engineering bar as any other production you enjoy owning problems end-to-end writing services that people rely on and collaborating across disciplines wed love to talk to you.

Were seeking a versatile technically strong Backend Software Engineer to design build and own backend infrastructure for imaging engineering and quality workflows across Camera Photos and Image Quality building and operating Python REST API services designing data models for enormous volumes of image and metadata records and running and scaling asynchronous compute jobs including the serving infrastructure for our AI/ML models. The ideal candidate has a solid grasp of distributed-systems fundamentals and is comfortable owning a service from API design through production operation writing code with an eye toward maintainability correctness and long-term operability and is equally at home designing a new service debugging a tricky async job standing up model-serving infrastructure or sitting with a partner team to understand what they actually need. You hold AI-powered features to the same engineering standards as any other production code and you treat cross-functional communication as a core part of the job.

Design deploy and operate model-serving infrastructure for LLM integrations agentic workflows and vision pipelines including hosting versioning monitoring and cost/latency/accuracy tradeoffs across the service lifecycle not just calling hosted third-party prompt engineering strategies and retrieval-augmented systems (RAG) and the underlying vector storage/retrieval infrastructure that make internal image and metadata corpora accessible and actionable to partner integrate and maintain AI/ML models in production: monitoring for quality regression managing model versions and balancing cost latency and accuracy tradeoffs across the service and maintain Python REST API backends and data models/storage for large-scale image video and metadata catalogs including endpoints that kick off monitor and scale long-running asynchronous with engineers scientists and quality leads across Camera Photos and Image Quality to translate their workflows into reliable backend the reliability performance and observability of services other teams depend on; contribute to technical design code review and cross-team planning.n

BS in Computer Science Computer Engineering or equivalent experience.n4 years of professional software engineering experience shipping production backend proficiency in Python with a track record of owning production backend services end to understanding of REST API design and experience building and operating production REST services at experience hosting and serving AI/ML models (LLMs vision models or similar) in production including infrastructure for inference scaling and monitoring not just integrating third-party hosted knowledge of asynchronous job execution patterns (background workers task queues or similar) for long-running computations and experience scaling these systems under understanding of distributed-systems fundamentals: consistency coordination failure handling and tradeoffs between understanding of software engineering fundamentals: data modeling API design testing debugging and code written and verbal communication skills with a demonstrated ability to work effectively with partners outside of engineering.

Hands-on experience with specific self-hosted GPU inference frameworks (e.g. vLLM Triton Ray Serve or similar) at production scale beyond the general hosting/serving experience required building production features with LLM APIs (e.g. OpenAI Anthropic or on-device models) including prompt design context window management output validation and graceful with multimodal or computer vision models applied to image analysis quality assessment or visual data retrieval with an understanding of where these models succeed and fail in with vector databases or semantic search (e.g. pgvector Pinecone Weaviate) for unstructured or high-dimensional data retrieval of MLOps principles: model deployment pipelines versioning strategies evaluation frameworks A/B testing for AI features and production monitoring for model quality and of bias and fairness considerations in AI systems particularly in visual domains including diverse evaluation datasets inclusive quality benchmarks and responsible deployment -on operational experience with container orchestration (e.g. Kubernetes) and infrastructure-as-code for distributed systems beyond the conceptual fundamentals required with Solr (or other search platforms such as Elasticsearch) for indexing and querying large with Redis whether as a cache message broker or coordination working with image data metadata pipelines or scientific/engineering and adaptable in a fast-paced environment with shifting priorities and multiple stakeholders.