About this opportunity
Blackhornvc lists this Head of Research opportunity in san francisco, California. Review the employer’s description below for duties, qualifications and application requirements.
Job description
Why This Role, Why Now
We are building the best inference platform for multi-agentic workflows. That platform spans four things: Design, Deploy, Run, and Optimize.
Today our differentiation sits in Deploy, where we have technology that scales agentic workflows better than anything else available. Design, Run, and Optimize are where the next advantage has to come from. We have working ideas on Run and Optimize, and nothing settled on Design.
We are ready for someone to own that research full time.
About the Role
We are looking for a Head of Research to lead CanyonLabs, working directly with Canyon Code's Chief Scientist, Chief Architect, and CEO. You will set and run the research agenda, build and supervise a programme of academic residents, and publish papers and open-source work alongside them. Equally, you will fold that research back into the product line. Both halves matter: research that never ships is a cost centre, and product that never advances the science is a commodity. This is a hands-on role.
You're a fit if you…
A PhD in Computer Science, Machine Learning, EE, or a related field, or an equivalent record of published research.
First-author publications at venues such as NeurIPS, ICML, ICLR, MLSys, OSDI, SOSP, or NSDI.
Strong preference for candidates who have done this job before: leading a research function at a hyperscaler, a frontier lab, or a research institute.
Depth in one or more of our research directions: recursive self-improvement of models and workflows, model fine-tuning, or kernel-level optimization for inference.
You have supervised academic residents, interns, or PhD students, and shipped papers with them.
You have taken research into production. You know which results survive contact with a real workload and which do not.
You can still build. You write the prototype that tests the idea rather than waiting for someone else to.
Clear written communication: you can turn research-grade ideas into papers, docs, and product language.
Self-driven; thrives in ill-defined, ambiguous, early-stage work.
Hands-on with multi-agentic systems built on LangGraph, AutoGen, CrewAI, Semantic Kernel, or custom agent stacks.Plus
Experience with model-serving platforms such as vLLM, SGLang, Ray, or NVIDIA Dynamo.Plus
An established open-source track record.Plus
Why Join
Real ownership
Competitive equity and compensation, and the impact to match. You are not joining a research org, you are creating one.
Lead CanyonLabs
The agenda, the academic resident programme, the publications, and the open-source output are yours to define and run.
Publish and ship
We expect the work to appear at top venues and in the product. Those are not competing goals here, they are the same goal.
Research with a production feedback loop
Every workflow running on CanyonOS generates real runtime data. You are not working from public benchmarks, you are working from how agentic applications actually behave in production.
Work alongside the founding team
Direct collaboration with our Chief Scientist, Chief Architect, and CEO, who bring deep experience in ML systems research and AI infrastructure platforms.
#J-18808-Ljbffr
Worksite address
san francisco, CA, 94199, US
Who can apply
Review the original listing for work authorization, qualifications and employer requirements.