Mlops

I’d love some career advice from people who’ve been in similar roles. I’ve been in MLOps for about 4–5 years, and most of my work has been pretty ops-heavy: Kubernetes, AWS, GKE, GPU debugging, CUDA/driver compatibility, and lately more agentic/AI infrastructure work like researching MCP gateways and MCP servers. Even though I’ve been part of a Machine Learning team, I’ve mostly stayed on the operations/infrastructure side. I originally wanted that setup because I hoped it would keep me close to ML research and applied ML, but in practice I don’t get many opportunities to work on those areas. Most of my time goes toward supporting ML engineers with ops and platform issues. So my experience is strong in areas like: * production reliability * deployment maturity * infra debugging * GPU/platform knowledge * scaling and cost control But I have much less hands-on exposure to: * applied ML * evaluation/benchmarking * prompt/context engineering * model behavior analysis Now I’ve been given the option to move more formally into a Cloud/DevOps team, and I’m trying to think long term. Given where AI seems to be heading — more agentic systems, infrastructure/platform work, and less emphasis on doing in-house model research because frontier models are increasingly available from large vendors — what do you think is the better path for career growth and job security? Would you stay closer to the ML org even if your work is mostly ops, or move fully into Cloud/DevOps / platform engineering and lean into that lane? I’d especially love to hear from people working in MLOps, applied ML, AI platform, or infra.