Aiml - Sr Backend Engineer, Data and ML Innovation

Apple Apple · Big Tech · Seattle, WA · Machine Learning and AI

This role focuses on building scalable backend pipelines and services for training data used in Foundation Models and Apple Intelligence features. Responsibilities include converting raw data into training formats, processing and filtering large datasets, and developing APIs for data access. The role also involves large-scale inferences for pre-training and post-training using fine-tuned LLMs.

What you'd actually do

  1. Training data pipeline development. Convert raw data into format acceptable by training jobs on GCP and AWS. Leverage internal and open-sourced training modules.
  2. Large scale inferences: Leverage internal and open-sourced inference stack to generate inferences with fine-tuned LLMs on massive amounts of data, for pre-train and post-training
  3. Data processing and data filtering: Have the ability to efficiently process and filter very large amounts of data, often times messy.
  4. Scalable web services backend and APIs to support data access and data inspection tools

Skills

Required

  • BS in Computer Science or Equivalent
  • 10+ years of programming experience in Python
  • Extensive experiences in concurrency and parallelism, functional programming, decorators
  • Proficient in REST API, Redis, VectorDB or other large scale data storage systems
  • Solid foundational programming skills (algorithms, data structures, OOP, etc)
  • Experience designing, writing, reviewing, testing and delivering software for applications and systems at scale

Nice to have

  • Familiarity with streaming-processing (Kafka)
  • Familiarity with a variety of build tools (Jenkins, Maven, Docker, Gradle)
  • Experience providing architecture and design guidance
  • A good communicator with clear and concise, active listening and empathy skills
  • Are self-motivated and curious. Strive to continually learn on the job.
  • Have demonstrated creative and critical thinking with an innate drive to improve how things work. Have a high tolerance for ambiguity.

What the JD emphasized

  • track record in building scalable pipelines and services
  • hundreds of millions of customers
  • billions of their interactions
  • massive amounts of data

Other signals

  • Foundation Models
  • Apple Intelligence
  • model hillclimbing
  • data augmentation
  • training efficiency