Staff Machine Learning Engineer

Unity Technologies SF; Mountain View (CA); On-site; USD 172,000 - 284,000; Full time

Benefits offered by this job

Equity awards

Annual discretionary bonuses

Sales commissions

Office snacks

Mental Health and Wellbeing programs

Employee Resource Groups

Training and development programs

Job summary

Unity Technologies SF is hiring a Staff Machine Learning Engineer to build production-grade AI for game experiences with a focus on computer vision and multi-modal modeling.

You will set the technical vision for CV and multi-modal models, design and deploy models for image and video understanding, and lead multi-modal inference systems across cloud and on-device targets.

Qualifications

• 6+ years of ML engineering focusing on computer vision and/or multi-modal modeling.
• Experience with transformer-based and diffusion-based vision models; e.g., ViT, CLIP, Stable Diffusion.

Responsibilities

• Set technical vision and roadmap for CV and multi-modal AI models.
• Design models for image/video understanding and generation (segmentation, detection, dense prediction).
• Develop multi-modal reasoning over images, text, and 3D inputs.
• Balance quality, latency, and cost across cloud, server, and on-device targets.
• Lead research-to-production delivery: training, fine-tuning, distillation, and serving.
• Collaborate with research scientists to translate architectures into deployable code.
• Build scalable multi-modal inference systems processing images, video, text, and metadata.
• Monitor breakthroughs in vision-language pretraining and efficient diffusion techniques.

Skills

Computer vision

Multi-modal modeling

Python

PyTorch

Transformer architectures

Diffusion models

Vision-language alignment

Reinforcement of on-device inference

Tools

TensorRT

ONNX Runtime

CoreML

TFLite

FlashAttention

ViT

CLIP

Stable Diffusion

DETR

SAM

Job description

Unity Technologies SF is hiring a Staff Machine Learning Engineer to build production-grade AI for game experiences with a focus on computer vision and multi-modal modeling.

Responsibilities

• Set technical vision and roadmap for computer vision and multi-modal AI models, covering transformers, diffusion models, vision-language models, and JEPA-style generative architectures
• Design and implement models for image and video understanding and generation, including segmentation, detection, and dense prediction
• Develop multi-modal reasoning over images, text, and 3D inputs
• Make architecture, training, data pipeline, and evaluation trade-offs balancing quality, capability, latency, and cost across cloud, server, and on-device targets
• Drive research-to-production delivery: training, fine-tuning, distillation, export, and serving for deployment scenarios from cloud GPUs to efficient on-device inference
• Partner with research scientists to translate novel CV and multi-modal architectures into deployable, well-engineered implementations
• Build scalable multi-modal inference systems that ingest diverse inputs (images, video, text, primitives, and metadata) and produce outputs ranging from semantic predictions to pixel-level generation
• Monitor and adopt field breakthroughs including vision-language pretraining and alignment, efficient diffusion approaches (consistency models, flow matching), efficient attention (FlashAttention, linear-attention variants), and vision tokenization/representation learning
• Where needed for latency or device constraints, apply compression and optimization such as compression, quantization, pruning, and knowledge distillation, and integrate runtimes like TensorRT, ONNX Runtime, CoreML, and TFLite
• Lead and mentor ML engineers, establishing engineering best practices, code review standards, and rigorous benchmarking and evaluation methodology
• Collaborate with research, platform engineering, product managers, and runtime teams to align ML capabilities to product roadmaps and target-platform constraints
• Define and enforce measurement practices using KPIs for model quality, accuracy, latency, memory, and cost

Requirements

• 6+ years of ML engineering with strong depth in computer vision and/or multi-modal modeling
• Production experience with transformer-based and diffusion-based vision models (examples: ViT, CLIP/SigLIP-style encoders, Stable Diffusion, DETR/SAM-style architectures)
• End-to-end model lifecycle experience including data curation, training and fine-tuning, evaluation, and serving at scale
• Familiarity with efficient attention, diffusion samplers, multi-modal fusion, and vision-language alignment methods
• Strong Python skills and modern deep-learning tooling such as PyTorch, plus solid software engineering fundamentals
• Proven technical leadership: setting direction, influencing cross-functional partners, and growing engineers

Technologies

• Python, PyTorch
• TensorRT, ONNX Runtime, CoreML, TFLite
• FlashAttention
• ViT, CLIP, SigLIP
• Stable Diffusion
• DETR, SAM

Benefits

• 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

Additional information

• Location: Mountain View, CA (onsite)
• Salary (USD per year): USD 172,200 - 283,900
• Zone A: $218,400 - $283,900
• Zone B: $194,100 - $252,300
• Zone C: $172,200 - $223,900
• Beyond base salary, the role may be eligible for equity awards and participation in company incentive plans (including annual discretionary bonuses or sales commissions)
• Final offer depends on geographic location, relevant experience, professional background, and skill set

You might also have

• Experience with world-model, video-generation, or neural rendering pipelines (NeRF, 3DGS, or similar)
• Experience deploying models to constrained or on-device targets, including quantization (INT8/INT4/FP16), pruning, distillation, and runtimes such as CoreML, TFLite, ONNX
• Familiarity with mobile SoC accelerators (Apple Neural Engine, Qualcomm Hexagon/Adreno, ARM Mali) or compiler stacks such as MLIR, TVM, or XLA
• Contributions to open-source ML frameworks or peer-reviewed CV/ML research publications
• Background in real-time graphics or game engine pipelines (Metal, Vulkan, OpenGL ES)