About the job
This position sits within Nebius Token Factory, our serverless platform for running and customizing open-source LLMs in production. Token Factory allows for serverless inference and fine-tuning backed by in-house optimizations like custom speculative decoding, quantization, cache-aware routing and dedicated endpoints. Customers come to us to move from prototype to scaled production without the cost and complexity of building and tuning their own inference stack.
Our Solutions Architects own the technical delivery of customer engagements: deploying open-source models, tuning the serving stack, benchmarking against the customer's success criteria, and carrying the technical relationship through to production. Sales Engineers qualify and scope the opportunity; SAs execute it. Technical Account Managers take it from production onward.
We're looking for a Manager, ML Solutions Architecture to lead our US regional SA teams. Your scope is PoC delivery and post-sales technical support: the people who do it, the standard they do it to, and the operating cadence that keeps it predictable. You will report to the Head of Solutions Architecture and partner with a peer manager in the other region.
Responsibilities
Manage a team of 4 Solutions Architects, with continued growth planned: 1:1s, goal setting, performance reviews, promotion cases, and individual growth plans
Be accountable for your team's delivery outcomes: time from PoC kick-off to first optimized dedicated endpoint, success-criteria hit rate, and the quality of the technical relationship after the customer goes to production
Maintain and extend the team's documentation: responsibilities, runbooks, guides, onboarding, definitions of done, engagement closure templates
Convert recurring customer pain into prioritized platform work, and represent the customer's technical reality in roadmap discussions
Hold the scoping-to-execution boundary: push back on under-scoped engagements, and feed feasibility signal back upstream
Make production handoffs uneventful, and keep post-sales technical requests moving
Qualifications
Minimum
3+ years managing technical teams, including performance management and difficult conversations
Experience managing a customer-facing team
Strong ML knowledge: LLM architectures, fine-tuning approaches (SFT/LoRA, RL-based), evaluation design, and a working command of inference internals — quantization, KV-cache management, batching, routing, speculative decoding — and of the frameworks the team works in (vLLM, SGLang, TensorRT-LLM)
Enough technical judgment to review someone else's benchmark and find the flaw in the methodology
Python strong enough to read and review your team's code
Excellent communication skills, with the ability to clearly explain technical concepts to diverse audiences
Genuine tolerance for operational work: documentation, process design, reporting, and the follow-through that makes them stick
Comfort operating with ambiguity across distributed teams and timezones, and a bias toward writing things down
Preferred
References from both former managers and former direct reports
Experience scaling a team through rapid growth (5 → 15+) without losing delivery quality
Prior experience in a customer-facing technical function at a cloud, inference, or AI infrastructure provider
Experience defining process and documentation for a team that had none
Hands-on background running LLMs in production and debugging inference workloads at the framework level
Work with multimodal AI models (vision-language, speech)
Proficiency with DevOps tooling (Docker, Kubernetes) and infrastructure-as-code