Turn research on models, inference, and hardware into work that helps engineers understand the stack they rely on.
Why this role exists
Software engineers increasingly rely on models, inference systems, and hardware for almost all of their day-to-day work, but rarely get a clear view of how those systems work. We think they should. Knowledge is power.
Our research team produces whitepapers about the inference stack. You will be the translation layer between this work and regular full-stack software engineers who want a deeper understanding of what they are building on top of.
What you will do
- Work with our researchers as their work develops, not only after a paper is finished.
- Turn our whitepapers into articles that teach full-stack engineers the concepts, evidence, and practical consequences of the research.
- Read new model papers, then explain the architecture, training process, evaluations, strengths, weaknesses, and implications for inference.
- Build small experiments, benchmarks, diagrams, and code examples when they make an explanation more useful or more honest.
- Develop a distinct technical reputation for Gradiated that engineers trust.
What we are looking for
- You have real software engineering experience and can investigate a technical claim in papers, code, traces, and benchmarks rather than repeating its summary.
- You write clearly and precisely, and you care about teaching rather than sounding clever.
- You can explain technical depth without removing the caveats, evidence, or uncertainty that make it trustworthy.
- You are curious enough to follow a question across models, training, inference software, systems, and hardware.
- You know how to interview technical people, challenge assumptions, and find the question that unlocks an explanation.
- You have writing, teaching, or explanatory work that shows how you think. It does not need to have been your job title.
This role is not for you if
- You don't already read lots of whitepapers.
- You're scared about learning about hardware if you're not in that world already.
- You haven't truly been an engineer building production systems before.
- You need lots of handholding.
#J-18808-Ljbffr
