Senior Applied AI Solutions Engineer

Nebius B.V.; Deutschland; Hybrid; EUR 90.000 - 130.000; Vollzeit

Benefits dieser Stelle

Competitive salary

Professional growth

Flexible working

Collaborative environment

Zusammenfassung

Nebius is expanding across serverless, ML platforms, and enterprise AI. This role sits at the intersection of deep ML engineering and product impact, spending roughly half time with customers to move POC to production and deliver hands-on onboarding.

The other half is spent building demos and prototypes, exploring emerging techniques, and translating findings into concrete product directions that accelerate adoption and value for enterprise customers.

Qualifikationen

• Experience fine-tuning large models and optimizing training processes.

Aufgaben

• Build prototypes and demos across the product portfolio — serverless inference, databases, MLflow, MLOps, and vertical use cases in Physical AI and HCLS — that become assets for sales, product, and engineering teams
• Support new customers hands-on through POC design, technical onboarding, and validation; act as the bridge between their ML team and the platform during the critical first months
• Go deep on emerging applied AI — new training techniques, inference optimizations, agentic architectures, new frameworks — and turn findings into working prototypes, writeups, and product recommendations
• Feed the product roadmap with specific, grounded feedback; be the voice of
• Develop reusable technical assets — notebooks, reference architectures, benchmark results — that reduce onboarding friction at scale

Kenntnisse

Large model fine-tuning

Distributed training debugging

Production ML pipelines

GPU inference optimization

Customer-facing ML work

Tools

PyTorch

HuggingFace

CUDA fundamentals

Kubernetes for ML

MLflow

Vector databases

Jobbeschreibung

The role

AI is moving faster than any single product team can track. Nebius is expanding across serverless, databases, MLflow, MLOps, Physical AI, and HCLS — and customers arriving with complex, real-world ML workloads need more than documentation. This role exists to close that gap: someone who can prototype what's possible, accelerate customers through their first 90 days, and feed hard-won field insight back into the product roadmap. This role sits at the intersection of deep ML engineering and product impact. You'll spend roughly half your time in the field — helping new customers move from POC to production, running technical onboarding, and working hands-on through their ML stack. The other half you'll spend building — prototyping applied AI use cases that show what's possible on the platform, going deep on emerging techniques before they're mainstream, and turning that expertise into concrete product direction. This is not a presales role. You get your hands dirty every day.

What success looks like in 12 months

• The product and sales teams have a library of working, polished demos they reach for on calls
• Enterprise customers you've touched have meaningfully faster time-to-value than those you haven't
• At least 2–3 product changes were shipped because of feedback you originated
• The team understands where applied AI is heading 6–12 months from now, partly because you told them

Your responsibilities will include:

• Build prototypes and demos across the product portfolio — serverless inference, databases, MLflow, MLOps, and vertical use cases in Physical AI and HCLS — that become assets for sales, product, and engineering teams
• Support new customers hands-on through POC design, technical onboarding, and validation; act as the bridge between their ML team and the platform during the critical first months
• Go deep on emerging applied AI — new training techniques, inference optimizations, agentic architectures, new frameworks — and turn findings into working prototypes, writeups, and product recommendations
• Feed the product roadmap with specific, grounded feedback; be the voice of "here's what broke in three customer POCs last month and here's what needs to change"
• Develop reusable technical assets — notebooks, reference architectures, benchmark results — that reduce onboarding friction at scale

We expect you to have:

• You’ve fine-tuned large models, debugged distributed training jobs, built production RAG or agentic pipelines, and optimized inference on GPU infrastructure — not just read about it
• You're fluent in the modern ML stack: PyTorch, HuggingFace, CUDA fundamentals, Kubernetes for ML, MLflow or equivalent, vector databases
• You've worked with enterprise ML teams — whether as a solutions engineer, customer engineer, or an ML engineer who collaborated closely with customers
• You read papers and implement them — not for credit, but because it's how you stay sharp
• You communicate with calibration: you can explain activation checkpointing tradeoffs to an ML engineer in the morning and the cost implication to a CTO in the afternoon

It will be an added bonus if you have:

• Experience in any of our vertical domains: Physical AI / robotics / simulation, HCLS (drug discovery, medical imaging, clinical NLP), or enterprise AI application development
• Familiarity with MLOps at scale (Kubeflow, Metaflow, Argo, Ray)
• Prior work at a cloud provider or AI infrastructure company
• You've shared technical work publicly — notebooks, talks, blog posts that people actually use

Who thrives here

You’ll thrive here if you're energized by variety — one day deep in a customer's MLOps stack, the next building a demo from scratch. You want your technical depth to influence product decisions, not just close deals.

What we offer

• Competitive salary and comprehensive benefits package.
• Opportunities for professional growth within Nebius.
• Flexible working arrangements.
• A dynamic and collaborative work environment that values initiative and innovation.

We're growing and expanding our products every day. If you're up to the challenge and are excited about AI and ML as much as we are, join us! You've fine-tuned large models, debugged distributed training jobs, built production RAG or agentic pipelines, and optimized inference on GPU infrastructure — not just read about it, You're fluent in the modern ML stack: PyTorch, HuggingFace, CUDA fundamentals, Kubernetes for ML, MLflow or equivalent, vector databases, You've worked with enterprise ML teams — whether as a solutions engineer, customer engineer, or an ML engineer who collaborated closely with customers, You read papers and implement them — not for credit, but because it's how you stay sharp, You communicate with calibration: you can explain activation checkpointing tradeoffs to an ML engineer in the morning and the cost implication to a CTO in the afternoon