Hybrid-Job

Hybrid / Homeoffice möglich

Senior AI Research Engineer - GPU Systems & Kernels

Datadog

Paris, France, Paris · FrankreichVeröffentlicht am 09. Nov. 2026Aktiv bestätigt am 16. Nov. 2026
Technology

Job-Briefing

Aus der Anzeige gelesen und mit unserem Stellenindex abgeglichen

Büro in
Paris, France, Paris
Arbeitsmodell
Hybrid, mit Präsenzanteil
Gehalt
Nicht angegeben
Level
Senior

Die Rolle in 30 Sekunden

Aus dem Englischen

Als Senior Research Engineer im Frontier-Foundations-Team machst du das Training und die Inferenz großer Modelle so schnell und effizient wie möglich. Die Rolle richtet sich an erfahrene Systems Engineers mit Expertise in GPU-Kernels, Kommunikation oder Trainingsinfrastruktur.

Deine Aufgaben

  • GPU-Kernels für wichtige Abläufe beim Training und bei der Inferenz von LLMs schreiben, optimieren und debuggen
  • Leistung auf aktuellen NVIDIA-GPUs optimieren und Speicher- sowie Kommunikationsengpässe reduzieren
  • Verteilte Trainings- und Rollout-Workloads profilieren und Engpässe sowie Optimierungspotenzial ermitteln
  • Benchmarks, Profiling-Tools und Regressionstests entwickeln
  • Mit Research Scientists effiziente Implementierungen neuer Modellarchitekturen erarbeiten
  • CUDA
  • Triton
  • NVIDIA
  • C++
  • Rust
  • Python
  • Nsight
  • NCCL
  • NVSHMEM
  • CUTLASS

Das wird erwartet

Fachlich gefordert

  • Fundierte Kenntnisse in mindestens einem Bereich: GPU-Kernel-Entwicklung, Kommunikation mit niedriger Latenz oder Trainingsinfrastruktur im
  • Hardwarebeschränkungen wie Speicher, Bandbreite und Verbindungstopologie analysieren und Engpässe verifizieren können
  • Sicher im Umgang mit einer Systemsprache wie C++, Rust oder CUDA C sowie mit Python
  • Starke Debugging- und Profiling-Fähigkeiten, auch mit Tools wie Nsight
  • Erfahrung mit Mixture-of-Experts-Modellen, Expert Parallelism oder Sparse Model Inference ist ein Plus

Haltung

  • Technische und nicht-technische Zielgruppen klar über Design- und Performance-Kompromisse informieren
  • Leidenschaft für Technologie und Motivation, die eigenen Fähigkeiten auszubauen

Gut zu wissen

Bei uns seit
9. November 2026
Arbeitgeber
51 weitere offene Stellen bei Datadog in unserem Index. Davon 4 vollständig remote.

Bevor du dich bewirbst

Diese Begriffe stehen wörtlich in der Anzeige. Wenn du sie beherrschst, gehören sie auch in deinen Lebenslauf.

  • GPU kernel development and optimization
  • low-latency communication and collective operations
  • large-scale training infrastructure
  • strong debugging and profiling skills
  • Mixture-of-Experts models
  • expert parallelism
  • Triton
  • distributed training frameworks
Passt die Stelle zu dir? Frag den KI-Berater

Automatisch erstellt von nomado24 aus der Anzeige und unserem Index. Keine Angabe des Arbeitgebers. Maßgeblich ist die Originalanzeige.

Stellenbeschreibung

As a Research Engineer on the Frontier Foundations team, you will partner with Research Scientists to make large-scale model training and inference as fast and efficient as the hardware allows. You will work at the lowest level of our stack, writing and tuning GPU kernels and removing the memory and communication bottlenecks that limit large Mixture-of-Experts models on the latest GPUs.

Building on our track record of AI-powered solutions (e.g., Bits AI, Bits Evolve, and our time series foundation model Toto), Datadog AI Research (DAIR) tackles high-risk, high-reward problems grounded in real-world challenges in cloud observability and security.

Frontier Foundations builds and maintains the high-performance infrastructure that lets us train and serve large language models at scale. It supports both of our research areas:

Trained Agents for Observability - Post-training models to operate autonomously across Datadog's domain. SRE incident response is our first target, with a clear path to code repair, security response, and infrastructure optimization. We build the simulation environments, RL training loops, and evaluation infrastructure needed to train agents that match or surpass frontier models at a fraction of the cost.

What You'll Do:

  • Write, tune, and debug GPU kernels (CUDA, Triton, or similar) for the operations that dominate LLM training and inference.
  • Optimize for the latest NVIDIA GPU generations, including memory hierarchy, tensor cores, and low-precision formats such as FP8 and FP4
  • Reduce communication cost in distributed workloads (all-to-all, all-reduce) and overlap communication with compute across NVLink and across nodes
  • Profile end-to-end training and rollout workloads to identify whether they are compute-, memory-, or communication-bound, and quantify the headroom
  • Build benchmarks, profiling tools, and regression tests that guide optimization decisions
  • Collaborate with Research Scientists to turn new model architectures into efficient implementations
  • Explore emerging techniques such as low-precision training and serving, speculative decoding, and fused or persistent kernels

Who You Are:

  • You have depth in at least one of: GPU kernel development and optimization (CUDA or similar), low-latency communication and collective operations (e.g., RDMA, NCCL), or large-scale training infrastructure
  • You can reason about hardware constraints (memory, bandwidth, interconnect topology) and explain where a bottleneck is and how you would confirm it
  • You are proficient in a systems language (e.g., C++, Rust, or CUDA C) and in Python
  • You have strong debugging and profiling skills, including tools such as Nsight
  • You can explain design and performance trade-offs clearly to both technical and non-technical audiences

Bonus Points (any of the following):

  • You have experience with Mixture-of-Experts models, expert parallelism, or sparse model inference
  • You are familiar with GPU communication libraries such as NCCL or NVSHMEM, and with compute/communication overlap
  • You have experience with Triton, CUTLASS, or similar kernel frameworks
  • You are familiar with distributed training frameworks such as Megatron-LM, TorchTitan, or DeepSpeed
  • You have built GPU-intensive systems in graphics, rendering, simulation, or scientific computing
  • You have contributed to open-source ML or systems infrastructure projects

Datadog values people from all walks of life. We understand not everyone will meet all the above qualifications on day one. That's okay. If you're passionate about technology and want to grow your skills, we encourage you to apply.

LI-Hybrid

About Datadog:

Datadog is the leading observability and security platform for the AI era, providing businesses with unified visibility across the technology stack to manage complexity at scale. It brings applications, infrastructure, data, models, and security into one place, using AI to detect and resolve issues before they impact customers. Trusted globally by Fortune 500 companies and high-growth AI leaders, Datadog enables businesses to move faster with clarity and confidence. Learn more about #DatadogLife on Instagram , LinkedIn, and Datadog Learning Center.

Equal Opportunity at Datadog:

Datadog is proud to offer equal employment opportunity to everyone regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender identity, veteran status, and other characteristics protected by law. We also consider qualified applicants regardless of criminal histories, consistent with legal requirements. Here are our Candidate Legal Notices for your reference.

Datadog endeavors to make our Careers Page accessible to all users. If you would like to contact us regarding the accessibility of our website or need assistance completing the application process, please complete this form . This form is for accommodation requests only and cannot be used to …

greenhouseDev Engen

Diese Stelle wird von einer externen Quelle bereitgestellt. Die Bewerbung erfolgt auf der Website der Quelle.

Neue Remote-Jobs wie „Research Engineer“ per E-Mail

Kostenlos und nur, wenn etwas Passendes erscheint.

Anmeldung per E-Mail bestätigen, jederzeit abbestellbar. Datenschutzerklärung.

Alle Remote Data Jobs

Nicht der richtige Job?

Treffer zu deinem Profil: zwei siehst du kostenlos, mit Nomado24 Pro bis zu 24. Pro ansehen