I'm Alexander Kovel — Machine Learning Engineer focused on LLMOps, model serving, and high-throughput inference.
Currently @ Wildberries — optimizing Triton Inference Server and GPU infrastructure (NVIDIA MIG) for computer vision workloads.
- Serving LLMs at scale with vLLM, SGLang, TGI, Triton
- Inference autoscaling on Knative + KServe, driven by Prometheus metrics
- Model deployment pipelines on Kubernetes via ArgoCD
- Optimizing CV services in Python and C++, accelerating GPU-bound workloads
