AI Platform Engineer
I build machines to do my job for me.
These days, that mostly means building the platform that lets other machines run.
I'm an AI Platform Engineer working on the infrastructure behind production AI systems. My work sits at the intersection of AI Platform Engineering and LLMOps, with a focus on LLM inference, model serving, Kubernetes/OpenShift, observability, performance benchmarking and production governance.
I work on deploying and operating language models and other AI workloads using KServe, vLLM, Helm and Argo CD, as well as OpenAI-compatible inference endpoints and the infrastructure around them. A big part of my work also involves benchmarking throughput, latency and concurrency, and making model deployments more repeatable.
What interests me most is what happens after a model has been trained:
- How do we serve it efficiently?
- What happens when concurrency increases?
- Where are the latency bottlenecks?
- Are we actually using our GPUs efficiently?
- How do we observe what is happening in production?
- And how do we make deploying the next model easier than the last one?
Before moving into AI Platform Engineering, I researched on foundation models, vision-language models, image super-resolution and denoising at A*STAR I²R.
I'm particularly interested in AI infrastructure, LLMOps, LLM inference, model serving, Kubernetes, GPU performance and AI observability.