Staff Machine Learning Engineer

Unity Technologies; United States; Remote; USD 218,000 - 284,000; Full time

Benefits offered by this job

Health insurance

Stock options

Retirement plan

Generous vacation days

Parental leave support

Job summary

Unity Technologies is hiring a Senior Machine Learning Engineer for On-Device & Mobile AI to bring state-of-the-art multi-modal models to run fast and small on mobile and constrained hardware. You will own export, quantization, kernel tuning, and a shipped feature inside the engine at interactive frame rates.

You’ll work across NPUs, mobile GPUs, and desktop GPUs to shape latency, memory, and battery impact.

Qualifications

• 5+ years in software/ML engineering with on-device/edge inference experience
• Deployment of transformer- and/or diffusion-based models on mobile/desktop hardware
• Hands-on with at least one major inference runtime (ONNX Runtime / ORT Web, CoreML, TFLite, ExecuTorch)
• Low-level performance work: read a frame capture and kernel trace, profiling tools
• Knowledge of quantization (INT4/INT8/FP16), weight sharing, pruning, distillation
• Understanding of target hardware: mobile SoCs and GPUs (Apple/Qualcomm/ARM)
• Strong Python; familiarity with browser-runtime languages (TypeScript/JavaScript, WGSL) is a plus
• Working fluency with deployed models and architecture reasoning
• Collaborative, reliable delivery and mentorship qualities

Responsibilities

• Inference & On-Device Optimization: model export, graph transformation, operator fusion, memory-layout planning, hardware-specific tuning across NPUs and GPUs
• Apply quantization, weight sharing, pruning, and distillation to meet latency/memory/power budgets and validate quality
• Low-level performance work: write/tune WebGPU compute shaders and native kernels; profile with browser/platform tools
• Build integration between ML runtime and game engine: real-time scheduling, memory pooling, zero-copy buffers, frame-budget management
• Develop tooling and CI benchmarks for on-device performance and SKUs
• Collaborate with research scientists to productionize CV and multi-modal architectures for on-device deployment
• Provide feedback to research on hardware constraints and op-support gaps
• Track breakthroughs in efficient inference and pragmatically apply movements that impact latency/memory/power
• Share knowledge, review code, mentor junior engineers

Skills

On-device inference

WebGPU/WGSL

Model optimization

Python tooling

C++/Obj-C/Swift

Tools

Metal

Vulkan

CUDA

D3D12

Job description

The opportunity

We are building the next generation of AI-driven game experiences, running generativemodels on-device, right where the players are — on phones, tablets, laptops, anddesktops. Our games run inside a modern, browser-native runtime (built on technologiessuch as WebGPU and WebNN), so the models that power these experiences must be deployedand accelerated entirely within that runtime. As a Senior Machine Learning Engineer forOn-Device & Mobile AI, you will take state-of-the-art multi-modal models — transformers,diffusion networks, and vision-language models (VLMs) — and make them run fast, small, and reliably on mobile and constrained hardware.

This is a deeply hands-on role. You will own the optimization and deployment of significantparts of the inference stack — from a trained checkpoint leaving research, through export,quantization, and kernel-level tuning, to a shipped feature running inside the engine atinteractive frame rates within a fixed memory and power budget. Your work directly shapesthe latency, quality, memory footprint, and battery profile of AI features experienced bybillions of players.

This role is for an engineer who is energized by the gap between a research model and ashipping, on-device product. If you enjoy profilers, frame captures, op-fusion, and shavingmilliseconds and megabytes, this is your role.

What you’ll be doing

• Inference & On-Device Optimization
• Own the optimization pipeline for the models you ship: model export, graphtransformation, operator fusion, memory-layout planning, and hardware-specific tuningacross NPU, mobile GPU, and desktop/laptop GPU.
• Apply quantization (INT4/INT8/FP16), weight sharing, structured/unstructured pruning,and knowledge distillation to hit hard latency, memory, and power budgets — and validatethem against quality bars.
• Do low-level performance work: write and tune WebGPU compute shaders (WGSL) and,where relevant, native kernels (Metal, Vulkan/SPIR-V compute, CUDA); profile with browserand platform tools (Chrome/Dawn GPU traces, PIX, Instruments/Metal System Trace,
• Snapdragon Profiler, Nsight, RenderDoc), and eliminate bottlenecks at the op andmemory-bandwidth level.
• Apply efficiency techniques — dynamic resolution, token reduction, cross-framecaching/reuse, reduced-step diffusion samplers — as engineering levers to meet budgetson target SKUs.
• Runtime & Systems Integration
• Work with WebGPU-targeted inference runtimes (ONNX Runtime Web, Transformers.js,WebLLM, TensorFlow.js) alongside native options (CoreML, ONNX Runtime, TFLite,ExecuTorch), and extend or build glue code where off-the-shelf options fall short of ourdiffusion and VLM workloads.
• Build parts of the integration between the ML runtime and the game engine: real-timescheduling, memory pooling, zero-copy buffer sharing between the inference and renderpaths, and frame-budget management alongside the renderer.
• Build supporting engineering for your components: model packaging and asset pipelines,on-device fallbacks and SKU-aware capability tiers, crash/quality telemetry, and automatedon-device benchmarking in CI.
• Research Productionization
• Partner with research scientists to turn novel CV and multi-modal architectures intoimplementations that are deployable, debuggable, and fast on device.
• Provide a feedback loop into research: surface hardware constraints, op-support gaps, andcost models early so model design and deployment converge.
• Track breakthroughs in efficient inference (efficient attention, distillation, reduced-stepdiffusion) and assess them pragmatically: what actually moves latency/memory/power onour target devices.
• Collaboration & Engineering Quality
• Contribute to engineering best practices, code-review standards, performance-regressiongates, and on-device benchmarking methodology.
• Support a culture of measurement: track KPIs for latency, quality, memory, and power forthe systems you work on, across the device matrix.
• Partner with platform engineers, product managers, and runtime teams to align your workwith device-SKU constraints and product roadmaps.
• Share knowledge and mentor junior and mid-level engineers through code review, pairing,and design discussion.

What we’re looking for

• 5+ years in software/ML engineering, with meaningful time focused on on-device / edgeinference or real-time, performance-critical systems.
• Production deployment of transformer- and/or diffusion-based models (e.g., ViT, StableDiffusion, CLIP/SigLIP-style encoders) on mobile, desktop, or embedded hardware —shipped, not just prototyped.
• Hands-on experience with at least one major inference runtime (ONNX Runtime / ORT Web,CoreML, TFLite, ExecuTorch) and a working understanding of operator fusion, memorylayout, and runtime scheduling.
• Low-level performance engineering: solid command of at least one GPU/compute API — WebGPU/WGSL, Metal, Vulkan, D3D12, or CUDA — and the profiling tools to go with it. You can read a frame capture and a kernel trace and reason about where the time and memorygo.
• Working knowledge of model-optimization techniques — quantization (INT4/INT8/FP16),weight sharing, pruning, and distillation — and the judgment to apply them to hit latencyand memory budgets. You use them effectively as engineering tools.
• Understanding of target hardware: mobile SoCs (Apple Neural Engine, QualcommHexagon/Adreno, ARM Mali) and/or desktop/laptop GPUs (Apple Silicon, NVIDIA, AMD,Intel).
• Strong Python for export pipelines and training-side tooling; familiarity with the corelanguages of a browser-native runtime (TypeScript/JavaScript, WGSL) is a plus.
• Working fluency with the models you deploy — enough to read an architecture, modify it fordeployment, and reason about accuracy trade-offs.
• A collaborative working style: clear communication, reliable delivery, and a willingness tosupport and learn from teammates.

You might also have

• Experience shipping world-model, neural-rendering, or real-time generative pipelinesNeRF, 3DGS, real-time diffusion, or similar) on device.
• Hands-on experience deploying models through WebGPU — e.g., ONNX Runtime WebWebGPU EP), Transformers.js, WebLLM, or TensorFlow.js — including writing/tuning WGSLcompute shaders.
• Game-engine or real-time-graphics background (Unity, Unreal, or a custom engine;Metal/Vulkan/D3D/OpenGL ES render pipelines) — especially integrating computeworkloads alongside a renderer.
• Contributions to open-source ML inference frameworks, runtimes, or GPU/compute librariesespecially in the WebGPU ecosystem (Dawn, wgpu, ORT Web, Transformers.js, WebLLM).
• Familiarity with compiler stacks (MLIR, TVM, IREE, XLA) for custom kernel generation andgraph optimization.
• Experience with on-device benchmarking infrastructure, performance-regression CI, anddevice-farm matrices.
• Proficiency in C++/Objective-C/Swift for runtime integration.

Additional information

• Relocation support is not available for this position
• Work visa/immigration sponsorship is not available for this position
• Salary range : 218 400.00 USD - 283 900. 00 USD

This range reflects the anticipated base salary for this position. Beyond base salary, this role may be eligible for equity awards and participation in our company incentive plans (such as annual discretionary bonuses or sales commissions). The final offer amount will depend on several factors, including geographic location and the candidate’s relevant experience, professional background, and skill set.

Benefits

At Unity, we want our team members to thrive. We offer a wide range of benefits designed to support well-being and work-life balance.

Please note: Benefits eligibility, specific offerings, and coverage vary based on the country and employment status.

• Comprehensive health, life, and disability insurance
• Commute subsidy
• Employee stock ownership
• Competitive retirement/pension plans
• Generous vacation and personal days
• Support for new parents through leave and family-care programs
• Office food snacks
• Mental Health and Wellbeing programs and support
• Employee Resource Groups
• Global Employee Assistance Program
• Training and development programs
• Volunteering and donation matching program

Life at Unity

Unity [NYSE: U] is the world’s leading game engine, powering play for more than 3 billion consumers each month. The top mobile games in the world, the most played PC indie titles, the most innovative console games, and virtually all of the top XR and Web Games are developed, deployed, and grown in Unity. Unity also enables teams across industries like automotive, manufacturing, and healthcare to design, simulate, and collaborate in 3D — closing the gap between ideas and reality. For more information, please visit www.unity.com.

Unity is a proud equal opportunity employer. We are committed to fostering an inclusive, innovative environment and celebrate our employees across age, race, color, ancestry, national origin, religion, disability, sex, gender identity or expression, sexual orientation, or any other protected status in accordance with applicable law. Our differences are strengths that enable us to support the growing and evolving needs of our customers, partners, and collaborators. If you have a disability that means there are preparations or accommodations we can make to help ensure you have a comfortable and positive interview experience, please fill out this form to let us know.

Pursuant to the San Francisco Fair Chance Ordinance, we will consider for employment qualified applicants with arrest and conviction records.