AIML - Distinguished Engineer, Foundation Model

Apple is revolutionizing artificial intelligence by developing sophisticated

foundation models that power intelligent features across our product ecosystem.

We are seeking a Distinguished Engineer to set the technical direction for the

systems that power our foundation model training - with an initial focus on the

inference engine that the foundation model team relies on for training, model

evaluation, and other needs in the model development loop.

This is a senior individual-contributor leadership role. You will be one of the

most senior technical voices for foundation model systems at Apple: defining

the vision, driving execution across many teams, and raising the bar for

engineering excellence. The role starts with the inference engine, but we

expect you to move fluidly into adjacent training systems areas as the needs of

the foundation model program evolve.

Description

engineered specifically for Apple silicon and for experiences that are private,

personal, and deeply integrated into the OS. Behind that modeling work sits a

demanding systems layer, and the inference engine is at its center.

Our inference engine is used by the foundation model team throughout the model

development lifecycle: generating and processing data and running rollouts for

training, powering large-scale model evaluation, and serving as an LLM judge

that scores and compares model outputs. These workloads are high throughput,

bursty, and tightly coupled to research iteration - the speed, efficiency, and

reliability of the engine directly set the pace at which the team can train and

improve models.

As a Distinguished Engineer, you will own the technical strategy for this

inference engine and the broader systems that support it. You will partner

closely with modeling and research teams to bring new capabilities into the

development loop, work across many internal teams with very different

requirements, and lead a diverse set of engineers in turning an ambitious

vision into shipped milestones. While inference is the initial focus, you will

also help shape adjacent areas - training infrastructure, data systems, and

evaluation. If you are drawn to hard systems problems where the research and

the infrastructure are inseparable, this is the role.

Responsibilities:

Set and drive the technical vision and roadmap for the foundation model

team's inference engine and the systems around it, used for training,

evaluation, and LLM-as-judge workloads.

Lead deep work on inference performance, efficiency, and reliability:

throughput and latency optimization, batching and scheduling, quantization,

speculative decoding, KV-cache management, memory and compute efficiency, and

hardware-aware optimization.

Architect inference systems that support a wide range of internal use cases -

data generation and rollouts for training, offline and large-scale evaluation,

and judge/reward scoring - across text, image, speech, and multi-modal models,

each with distinct throughput, cost, and quality constraints.

Extend your impact into adjacent systems areas - training infrastructure, data

pipelines, and evaluation harnesses.

Partner with many teams that depend on the engine, translating their diverse

needs into a coherent platform, clear interfaces, and a prioritized roadmap.

Work closely with ML researchers and modeling teams to co-design models and

systems, and to bring state-of-the-art techniques from prototype into the

development loop reliably.

Lead a diverse set of engineers across teams in setting direction and

executing against it; align stakeholders, resolve technical trade-offs, and

make the calls that keep large efforts moving.

Drive prioritization and milestone delivery across competing demands, balancing

near-term research needs against long-term platform investment.

Mentor and grow junior and senior engineers; establish engineering

standards, review designs, and multiply the impact of the organization.

Preferred Qualifications

Experience building or leading inference systems for large language models and

multi-modal foundation models at scale.

Experience with inference in training, evaluation, or reinforcement-learning

loops (e.g., large-scale rollouts, offline eval, or LLM-as-judge / reward

scoring).

Familiarity with Kubernetes, Docker, and cloud platforms (AWS, GCP, Azure),

and with distributed computing frameworks.

History of defining technical strategy that shaped an organization's or the

industry's direction.