Sr. Software Development Engineer (Applied ML)-Hardware Product Design
Apple
Software Engineering, Other Engineering, Product, Design, Data Science
Sunnyvale, CA, USA
USD 181,100-318,400 / year + Equity
Posted on Jun 9, 2026
We’re looking for a senior AI/ML engineer to do applied ML at the intersection of 3D geometry, manufacturing process, and the tacit expertise of the people who run that process. You’ll join the Data Science and Visualization (DataViz) team in Hardware Engineering at Apple, working day-to-day in close partnership with Apple’s Advanced Development Lab (ADL) to bring machine learning into the heart of their machining and prototyping workflows. Much of this work sits at the intersection of two things: CAD files that describe the parts ADL manufactures, and the subject matter expertise required to generate functional and beautiful parts. Your job will be to integrate ML and AI capabilities into the manual and routine parts of the process, and implement those solutions to a standard machinists can genuinely trust, so they can focus on the aspects that truly require their expertise.
As a senior member of this team, you will design, build, and own ML systems end-to-end for ADL’s machining, design-for-manufacturing, and related engineering workflows, including the architectural calls about which approach fits a given problem and when to retire one that isn’t scaling. You’ll work directly with the people running those workflows: understanding their constraints, building tools they trust, and iterating with tight feedback loops. You’ll choose the right tool for each job (classical statistics, classical ML, deep learning, generative AI, or pure algorithmic approaches) and make sure others understand your logic. The DataViz team is small. You’ll be the senior ML IC partnering with data scientists and visualization engineers on our side, with engineers and machinists on the ADL side, and with a partner engineering team that contributes to the broader system. Expect real autonomy on the architectural calls, and real accountability for whether the systems you ship still work six months later. We don’t expect any one candidate to bring every qualification below. What we care about most is the kind of thinking you bring to hard problems: clarity about what you do and don’t know, and the patience to work through ambiguity (and change your mind when the evidence asks you to). If that resonates, we’d love to hear from you.
- Own ML systems end-to-end (problem framing through deployment) for ADL’s machining, design-for-manufacturing, and related engineering workflows.
- Architect multi-component AI workflows: how models, agents, and rule-based components compose, where boundaries should sit, and how the pieces stay debuggable as the system evolves.
- Work with 3D geometric data (CAD files) including reasoning about geometry as a first-class input to models rather than a side channel.
- Select and apply the right modeling approach per problem.
- Establish evaluation and monitoring strategies that survive contact with messy real-world data, including offline benchmarks, automated checks, and human-in-the-loop review.
- Communicate trade-offs, system behavior, and limitations clearly to technical and non-technical audiences.
- Bachelor’s + 7 YOE, Master’s + 5 YOE, or PhD + 2 YOE (or equivalent professional experience) in CS, Math, Statistics, Physics, Engineering, Robotics, or a similar analytical field, with the bulk of those years building ML systems in production or applied settings.
- Strong Python skills and fluency with the standard ML stack. Practical fluency with using and evaluating modern foundation models (LLMs, VLMs) in production matters more than depth in any one training framework.
- Full-lifecycle ML experience covering problem framing, data work, training, evaluation, and iteration with real users, including the judgment to know when ML isn’t the right tool at all.
- Hands-on experience designing or extending agentic AI systems (multi-step, tool-using workflows where models plan, act, and recover) and the evaluation frameworks that keep them reliable — including ground truth that is contested, expensive, or partial, and small labeled datasets.
- Comfort collaborating with non-ML domain experts: drawing out tacit expertise that may never have been written down, translating their constraints into modeling decisions, and communicating results back in their terms.
- None of these are required, but any would strengthen your application. If you bring direct experience, please highlight it; if not but you’re excited to grow into them, we’d still encourage you to apply.
- Familiarity with 3D geometric data formats (STEP, mesh, BRep) or 3D libraries such as OpenCASCADE, Trimesh, or PyTorch3D.
- ML applied in engineering, manufacturing, machining, or other physical-world domains where models interact with hardware that has real-world constraints.
- Building reliable, user-facing features or workflows backed by LLMs, VLMs, or other GenAI models, particularly where visual or geometric inputs matter.
- Algorithmic depth relevant to manufacturing software: computational geometry, graph algorithms, constraint satisfaction, tool-path generation, or CAD/CAM internals.