Principal Engineer, Model Optimizations

You will own the model optimization discipline end to end: quantization strategy and its accuracy budget, kernel selection and authoring, attention and MoE execution, speculative decoding, and the parallelism layout that ties them together. You'll do it across model architectures that change every few months and across two vendor stacks with genuinely different performance characteristics—CUDA/Hopper/Blackwell on one side, ROCm/MI300-class and beyond on the other. The hard part is not making one model fast on one GPU. It's building the methodology, tooling, and upstream relationships that make every model fast on every GPU we buy, and doing it fast enough to launch a new model the week it drops.

What You'll Be Doing:

• Setting the technical strategy for model optimization across the fleet: which techniques we invest in, which we consume upstream, and how we decide
• Owning quantization end to end—FP8, FP4/MXFP4, INT8, weight-only schemes such as AWQ and GPTQ—including calibration methodology, the accuracy budget we're willing to spend, and the eval gates that enforce it
• Driving performance for modern architectures at the execution level: MoE routing and fused expert kernels, MLA and GQA attention variants, long-context and sliding-window attention, and the memory-movement patterns each implies
• Leading speculative decoding work (draft models, EAGLE/Medusa-class methods, n-gram and lookahead approaches) and the acceptance-rate tuning that determines whether it actually pays
• Writing and tuning kernels where upstream falls short—CUDA, Triton, CUTLASS, and their ROCm counterparts (HIP, Composable Kernel, hipBLASLt, AITER)—and knowing when not to
• Establishing how we choose parallelism layouts (TP, PP, EP, attention DP) per model, per GPU family, per traffic shape, and encoding that into something repeatable rather than tribal knowledge
• Building the benchmarking and regression infrastructure that makes optimization claims trustworthy: agent-shaped and long-context traffic distributions, TTFT/ITL/throughput measurement, and accuracy validation that runs before anything ships
• Making AMD a genuinely first-class target rather than a port, and driving the upstream work in vLLM, SGLang, and TensorRT-LLM required to get there
• Partnering with NVIDIA and AMD engineering teams on pre-silicon enablement, early-access hardware, and roadmap feedback
• Setting technical direction, mentoring senior and staff engineers, and representing DigitalOcean in upstream communities, at conferences, and in customer technical deep dives

What You'll Add to DigitalOcean:

• 12+ years in performance-critical systems, with substantial recent experience optimizing LLM inference in production
• Deep understanding of GPU architecture and the inference performance model: memory bandwidth vs. compute bounds, arithmetic intensity, kernel launch and scheduling overhead, and where prefill and decode differ
• Hands-on kernel-level experience on at least one vendor stack and the demonstrated ability and appetite to work across both—CUDA/CUTLASS/Triton and ROCm/HIP/CK
• Practical quantization expertise, including the judgment to know when a technique that benchmarks well will fail a customer's accuracy bar
• Familiarity with the internals of at least one major serving engine (vLLM, SGLang, TensorRT-LLM), sufficient to land non-trivial changes upstream
• Strong Python and C++/CUDA skills, and rigor in profiling with Nsight, rocprof, and equivalent tooling
• A measurement-first disposition: claims backed by reproducible benchmarks, with an understanding of how easily inference benchmarks mislead
• Excellent written and verbal communication, and experience leading cross-functional efforts spanning infrastructure, product, and customers

Bonus:

• Meaningful upstream contributions to vLLM, SGLang, TensorRT-LLM, PyTorch, Triton, or ROCm
• Experience bringing up a new accelerator family for production inference, including the unglamorous parts—numerics differences, missing kernels, immature libraries
• Experience with disaggregated prefill/decode serving and its interaction with parallelism and quantization choices
• Track record of same-week or day-zero enablement for newly released frontier open models
• Publications or patents in efficient inference, quantization, or GPU kernel design

Compensation Range:

• $249,600 - $312,000

*This is a hybrid role

JR: 2026-8158

#LI-Hybrid

Why You’ll Like Working for DigitalOcean

• We innovate with purpose. You’ll be a part of a cutting-edge technology company with an upward trajectory, who are proud to simplify cloud and AI so builders can spend more time creating software that changes the world. As a member of the team, you will be a Shark who thinks big, bold, and scrappy, like an owner with a bias for action and a powerful sense of responsibility for customers, products, employees, and decisions.
• We prioritize career development. At DO, you’ll do the best work of your career. You will work with some of the smartest and most interesting people in the industry. We are a high-performance organization that will always challenge you to think big. Our organizational development team will provide you with resources to ensure you keep growing. We provide employees with reimbursement for relevant conferences, training, and education. All employees have access to LinkedIn Learning's 10,000+ courses to support their continued growth and development.
• We care about your well-being. Regardless of your location, we will provide you with a competitive array of benefits to support you from our Employee Assistance Program to Local Employee Meetups to flexible time off policy, to name a few. While the philosophy around our benefits is the same worldwide, specific benefits may vary based on local regulations and preferences.
• We reward our employees. The salary range for this position is based on market data, relevant years of experience, and skills. You may qualify for a bonus in addition to base salary; bonus amounts are determined based on company and individual performance. We also provide equity compensation to eligible employees, including equity grants upon hire and the option to participate in our Employee Stock Purchase Program.
• DigitalOcean is an equal-opportunity employer. We do not discriminate on the basis of race, religion, color, ancestry, national origin, caste, sex, sexual orientation, gender, gender identity or expression, age, disability, medical condition, pregnancy, genetic makeup, marital status, or military service.

Application Limit: You may apply to a maximum of 3 positions within any 180-day period. This policy promotes better role-candidate matching and encourages thoughtful applications where your qualifications align most strongly.